From ben at ucc.gu.uwa.edu.au Wed Sep 1 07:55:04 1999 From: ben at ucc.gu.uwa.edu.au (Ben Rampling) Date: Wed Oct 27 01:17:13 2004 Subject: [tech] hdd4mako In-Reply-To: <19990831224117.F17478@mermaid.ucc.gu.uwa.edu.au> from Nick Bannon at "Aug 31, 99 10:41:18 pm" Message-ID: <199908312355.HAA14409@mermaid.ucc.gu.uwa.edu.au> > The suggestion was to stick a couple of the small SCSI drives from the > IBM gear into it with a SCSI card from Simon. (or I have some bootable > 8 bit TMC-850's in a pinch) They should be in the clubroom. Oh, I thought we actually had another solution rather than another "fuck off and let us spend money on beer". So we're going to switch to a set of drives that we're not even sure work, let alone sure if they're reliable? Aren't you tired of the way Committee is working, too, Nick? :) > I'm still in favour of picking up one or two new EIDE drives, but it > doesn't seem vitally urgent. It is urgent, because come Friday evening, Mako won't have a drive. Regards, Ben Rampling From fryers at ucc.gu.uwa.edu.au Wed Sep 1 10:51:54 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:14 2004 Subject: [tech] hdd4mako In-Reply-To: <199908312355.HAA14409@mermaid.ucc.gu.uwa.edu.au>; from Ben Rampling on Wed, Sep 01, 1999 at 07:55:04AM +0800 References: <19990831224117.F17478@mermaid.ucc.gu.uwa.edu.au> <199908312355.HAA14409@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990901105154.E14778@ucc.gu.uwa.edu.au> Bingle > A while ago Ben Rampling tapped: > > The suggestion was to stick a couple of the small SCSI drives from the > > IBM gear into it with a SCSI card from Simon. (or I have some bootable > > 8 bit TMC-850's in a pinch) They should be in the clubroom. > > Oh, I thought we actually had another solution rather than another "fuck off > and let us spend money on beer". So we're going to switch to a set of drives > that we're not even sure work, let alone sure if they're reliable? Aren't you > tired of the way Committee is working, too, Nick? :) They are not the drives I had in mind. There are a couple of drives sitting above utu in the UCC. Using a couple of them (You only need 400Mb so there should not be much problem) set up with / and swap on one disk and /usr on the other then there should not be much of a problem. Richard was running one of the drives with the SCSI controller I am bringing in on a FreeBSD box so there should not be any issues. It is not a case of UCC committee not spending money but finding an optimal solution to the problem. The drives are there - I am very supprised that you could not find them, which can provide a solution to your problem. > > I'm still in favour of picking up one or two new EIDE drives, but it > > doesn't seem vitally urgent. > > It is urgent, because come Friday evening, Mako won't have a drive. So, what services is Mako running that can not be transfered to another UCC machine? FMS? Web serving? See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From nick at ucc.gu.uwa.edu.au Wed Sep 1 11:48:17 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:15 2004 Subject: [tech] hdd4mako In-Reply-To: <19990901105154.E14778@ucc.gu.uwa.edu.au>; from Simon Fryer on Wed, Sep 01, 1999 at 10:51:54AM +0800 References: <19990831224117.F17478@mermaid.ucc.gu.uwa.edu.au> <199908312355.HAA14409@mermaid.ucc.gu.uwa.edu.au> <19990901105154.E14778@ucc.gu.uwa.edu.au> Message-ID: <19990901114816.L17478@mermaid.ucc.gu.uwa.edu.au> On Wed, Sep 01, 1999 at 10:51:54AM +0800, Simon Fryer wrote: > They are not the drives I had in mind. There are a couple of drives sitting > above utu in the UCC. Using a couple of them (You only need 400Mb so there [...] Aaaah. Details are good. [...] > So, what services is Mako running that can not be transfered to another UCC > machine? FMS? Web serving? Well... if minimising downtime is the aim, that narrows the options a little. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From gbowland at ucc.gu.uwa.edu.au Wed Sep 1 14:15:40 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:15 2004 Subject: [tech] Windowmanagers Message-ID: <19990901141539.A12159@mussel.ucc.gu.uwa.edu.au> The two SGIs that are functioning at the moment now have working copies of WindowMaker, because the default window manager was annoying me. It isn't the default window manager though, so you'll need a .xsession file to use it. The binary is /usr/freeware/bin/wmaker. -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From mustang at ucc.gu.uwa.edu.au Wed Sep 1 14:28:05 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:15 2004 Subject: [tech] Windowmanagers In-Reply-To: <19990901141539.A12159@mussel.ucc.gu.uwa.edu.au> from Grahame Bowland at "Sep 1, 99 02:15:40 pm" Message-ID: <199909010628.OAA16701@mermaid.ucc.gu.uwa.edu.au> > The two SGIs that are functioning at the moment now have working copies of WindowMaker, because the default window manager was annoying me. It isn't the default window manager though, so you'll need a .xsession file to use it. > > The binary is /usr/freeware/bin/wmaker. Good. Make it the default and I'll break your thumbs. 4Dwming happily, /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From maset at ucc.gu.uwa.edu.au Wed Sep 1 16:03:45 1999 From: maset at ucc.gu.uwa.edu.au (Anil Sharma) Date: Wed Oct 27 01:17:15 2004 Subject: [tech] Windowmanagers In-Reply-To: <19990901141539.A12159@mussel.ucc.gu.uwa.edu.au> Message-ID: Is that other hard drive on erwin yet? If so can I start exporting it and moving things like window managers and netscape to it? Maset the Grandiose. ----------------------------------------------- Without suffering, how can one appreciate happiness? And how would we mark the depths of our sorrow, without the light of hope? From ben at ucc.gu.uwa.edu.au Thu Sep 2 08:44:00 1999 From: ben at ucc.gu.uwa.edu.au (Ben Rampling) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] icmp Message-ID: <199909020044.IAA20382@mermaid.ucc.gu.uwa.edu.au> Sep 2 09:00:33 mako /kernel: icmp-response bandwidth limit 148/100 pps Is someone we know being a monkey, or is that someone else causing that? :) Regards, Ben Rampling From gbowland at ucc.gu.uwa.edu.au Thu Sep 2 08:57:33 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] icmp In-Reply-To: <199909020044.IAA20382@mermaid.ucc.gu.uwa.edu.au> References: <199909020044.IAA20382@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902085733.B16447@mussel.ucc.gu.uwa.edu.au> > On Sep 02, Ben Rampling scrawled : > Sep 2 09:00:33 mako /kernel: icmp-response bandwidth limit 148/100 pps > > Is someone we know being a monkey, or is that someone else causing that? :) I just strobed mako to see what was running - that might have done it. Sorry :) See you :) -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From alastair at ucc.gu.uwa.edu.au Thu Sep 2 10:35:04 1999 From: alastair at ucc.gu.uwa.edu.au (Alastair Irvine) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] Host key on mussel Message-ID: <19990902103504.A20681@ucc.gu.uwa.edu.au> Has somebody changed the ssh host key on mussel? When I tried to connect from mermaid, this is what I got: @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ @ WARNING: HOST IDENTIFICATION HAS CHANGED! @ @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY! Someone could be eavesdropping on you right now (man-in-the-middle attack)! It is also possible that the host key has just been changed. Please contact your system administrator. Add correct host key in /home/ucc/alastair/.ssh/known_hosts to get rid of this message. Agent forwarding is disabled to avoid attacks by corrupted servers. X11 forwarding is disabled to avoid attacks by corrupted servers. Are you sure you want to continue connecting (yes/no)? no -- ... When in doubt, Cheat! _______________________________________________________________________ | | | -=*Alastair Irvine*=- WWW page: http://members.xoom.com/alastair/ | | System Administrator Magic:tG player, coder, RPGer, net-nut | | e-mail: alastair@ucc.gu.uwa.edu.au or alastair.irvine@usa.net | |_______________________________________________________________________| From warrick at wyvern.com.au Thu Sep 2 10:42:02 1999 From: warrick at wyvern.com.au (Warrick Mitchell) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] Host key on mussel In-Reply-To: <19990902103504.A20681@ucc.gu.uwa.edu.au> References: <19990902103504.A20681@ucc.gu.uwa.edu.au> Message-ID: <19990902104202.A9746@wyvern.com.au> On Thu, Sep 02, 1999 at 10:35:04AM +0800, Alastair Irvine wrote: > Has somebody changed the ssh host key on mussel? When I tried to connect > from mermaid, this is what I got: > > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ > @ WARNING: HOST IDENTIFICATION HAS CHANGED! @ > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ Alaistair, yes it has changed, I think it may have to do with the fact someone did a upgrade on it recently, ie like over the weekend. Just edit your known hosts file and it will be sweet :) Seeya, Warrick From nick at ucc.gu.uwa.edu.au Thu Sep 2 11:14:38 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] Host key on mussel In-Reply-To: <19990902104202.A9746@wyvern.com.au>; from Warrick Mitchell on Thu, Sep 02, 1999 at 10:42:02AM +0800 References: <19990902103504.A20681@ucc.gu.uwa.edu.au> <19990902104202.A9746@wyvern.com.au> Message-ID: <19990902111438.V17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 10:42:02AM +0800, Warrick Mitchell wrote: > On Thu, Sep 02, 1999 at 10:35:04AM +0800, Alastair Irvine wrote: > > Has somebody changed the ssh host key on mussel? When I tried to connect > > from mermaid, this is what I got: > > > > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ > > @ WARNING: HOST IDENTIFICATION HAS CHANGED! @ > > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ > > Alaistair, yes it has changed, I think it may have to do with the fact someone did a upgrade on it recently, ie like over the weekend. Just edit your known hosts file and it will be sweet :) Maybe I'm forgetful, but I don't remember it changing for _me_, or even hearing about an upgrade. mussel: /etc>ls -l ssh* #11:08AM total 6 -rw-r--r-- 1 root wheel 1177 Jul 6 22:38 ssh_config -rw------- 1 root wheel 527 May 20 04:36 ssh_host_key ^^^^^^ -rw-r--r-- 1 root wheel 330 May 25 03:15 ssh_host_key.pub -rw------- 1 root wheel 512 Sep 2 10:10 ssh_random_seed -rw-r--r-- 1 root wheel 690 Aug 9 19:54 sshd_config If you last logged into mussel, without warnings, by ssh after May 20th, perhaps you _should_ be worried. It's odd that the ssh_host_key and ssh_host_key.pub don't match in date. As a general rule - even during upgrades you can avoid breaking the host key. I don't know whether this is strictly desirable or not. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From fryers at ucc.gu.uwa.edu.au Thu Sep 2 11:51:07 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:16 2004 Subject: [tech] HDD for Mako/Moray/Mermaid Message-ID: <19990902115107.I14778@ucc.gu.uwa.edu.au> Morning All I think I will add my cent or so on the technical discussion on the hard drive issue as I was responsible for Mel's knowledge at the committee meeting as to what was lying about. I probably should have replied to Ben's email earlier about the hard drive issue and actually turned up to the committee meeting but I had other commitments on tuesday afternoon. I have found a hard drive suitable for mako (>400MB), brought in the SCSI card and found a SCSI cable and refitted the termination resistors to the drive. The pile was sitting of the coffee table in UCC yesterday evening. It should still be there. The drive and controller have been used by Richard under FreeBSD so there should not be any problems with it in mako. To my knowledge, Mako is a 486 ISA based machine which does mostly network type things. The conclusion I draw from this is that the performance of Mako will not be affected be the (slower?) drive. The change from IDE to SCSI on mako whould be rather trivial I would imagine. Shut down all the services on mako, tar/ftp a copy of the file system to mussel or /home as needed. Replace hardware and boot using favouret OS that can format and write UFS file systems. ifconfig and ftp/untar the file system back again. Create /dev driectory. Rebuild kernel with scsi card drivers. Reboot. This sounds a lot like a starfish install apart from the rebuilding kernel bit! I did like Mustangs suggestion of getting another drive for Mermaid and dropping Mermaids drive into another machine, but I think there are more deserviong machines than Mako for this - moray for instance would probably be happier with a slightly larger drive. This is just my cent or two on the matter. Please feel free to flame my opions to death. The SCSI option was suggested as an optimal way of achieveing close to the same result with lower outlay from UCC. See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From luyer at ucc.gu.uwa.edu.au Thu Sep 2 13:29:19 1999 From: luyer at ucc.gu.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:17 2004 Subject: [tech] Switches and hubs... Message-ID: <199909020529.NAA20373@mussel.ucc.gu.uwa.edu.au> The second 10/100Mbps hub is now in the machine room (I believe Mustang has taken the first one, which was for him). I have turned the 10/100Mbps switch up the other way. To whoever placed it sitting on its intake fan - please think! David. From alastair at ucc.gu.uwa.edu.au Thu Sep 2 14:21:29 1999 From: alastair at ucc.gu.uwa.edu.au (Alastair Irvine) Date: Wed Oct 27 01:17:17 2004 Subject: Host key on *beowulf* [was: Re: [tech] Host key on mussel] In-Reply-To: <19990902111438.V17478@mermaid.ucc.gu.uwa.edu.au>; from Nick Bannon on Thu, Sep 02, 1999 at 11:14:38AM +0800 References: <19990902103504.A20681@ucc.gu.uwa.edu.au> <19990902104202.A9746@wyvern.com.au> <19990902111438.V17478@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902142129.F20681@ucc.gu.uwa.edu.au> On Thu, 02 September, 1999 at 11:14:38AM +0800, Nick Bannon wrote: > On Thu, Sep 02, 1999 at 10:42:02AM +0800, Warrick Mitchell wrote: > > On Thu, Sep 02, 1999 at 10:35:04AM +0800, Alastair Irvine wrote: > > > Has somebody changed the ssh host key on mussel? When I tried to connect > > > from mermaid, this is what I got: > > > > > > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ > > > @ WARNING: HOST IDENTIFICATION HAS CHANGED! @ > > > @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ > > > > Alaistair, yes it has changed, I think it may have to do with the fact someone did a upgrade on it recently, ie like over the weekend. Just edit your known hosts file and it will be sweet :) The host key on mussel has not changed. I screwed up; it was _beowulf_ I was trying to connect to. I usually use rsh; due to a thinko, I sshed to beowulf when I meant mussel. Using an existing rsh connection to beowulf (with which I had found xemacs to still not be installed), I catted /etc/ssh_host_key.pub and found that it is different from the beowulf line in my ~/.ssh/known_hosts . However, once I did ssh to beowulf successfully, I got this message: Warning: Remote host denied X11 forwarding, perhaps xauth program could not be run on the server side. [snip mussel details] > > As a general rule - even during upgrades you can avoid breaking the host > key. I don't know whether this is strictly desirable or not. Hmm... I guess JP didn't think about that. No problem. -- ... The wise learn more from fools than fools from the wise. _______________________________________________________________________ | | | -=*Alastair Irvine*=- WWW page: http://members.xoom.com/alastair/ | | System Administrator Magic:tG player, coder, RPGer, net-nut | | e-mail: alastair@ucc.gu.uwa.edu.au or alastair.irvine@usa.net | |_______________________________________________________________________| From mustang at ucc.gu.uwa.edu.au Thu Sep 2 14:52:11 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:17 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020529.NAA20373@mussel.ucc.gu.uwa.edu.au> from David Luyer at "Sep 2, 99 01:29:19 pm" Message-ID: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> > > The second 10/100Mbps hub is now in the machine room (I believe Mustang > has taken the first one, which was for him). Indeed I have. Sorry for not alerting people earlier. > I have turned the 10/100Mbps switch up the other way. To whoever placed > it sitting on its intake fan - please think! Hmm. Mtearle thinks its time for a reshuffle - think about how the network should be laid out WRT the switch & pass on your ideas. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From luyer at ucs.uwa.edu.au Thu Sep 2 14:57:01 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:17 2004 Subject: [tech] Switches and hubs... In-Reply-To: Your message of "Thu, 02 Sep 1999 14:52:11 +0800." <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> Message-ID: <199909020657.OAA05413@typhaon.ucs.uwa.edu.au> > Hmm. > Mtearle thinks its time for a reshuffle - think about how the network > should be laid out WRT the switch & pass on your ideas. The ascii art which I sent out before represents my opinion on how it should be done. David. From mustang at ucc.gu.uwa.edu.au Thu Sep 2 14:58:50 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:17 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020657.OAA05413@typhaon.ucs.uwa.edu.au> from David Luyer at "Sep 2, 99 02:57:01 pm" Message-ID: <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au> > > Hmm. > > Mtearle thinks its time for a reshuffle - think about how the network > > should be laid out WRT the switch & pass on your ideas. > > The ascii art which I sent out before represents my opinion on how it should > be done. I've got to admit that I missed it. Hence me asking. Got a pointer to it somewhere..? -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From luyer at ucs.uwa.edu.au Thu Sep 2 15:10:31 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:18 2004 Subject: [tech] Switches and hubs... In-Reply-To: Your message of "Thu, 02 Sep 1999 14:58:50 +0800." <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au> Message-ID: <199909020710.PAA05771@typhaon.ucs.uwa.edu.au> Perhaps someone else can re-post it, I don't know if I still have it. David. From mustang at ucc.gu.uwa.edu.au Thu Sep 2 15:11:35 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:18 2004 Subject: [tech] Mermaid re-animated Message-ID: <199909020711.PAA22105@mermaid.ucc.gu.uwa.edu.au> OK, so I got a faster CPU for it. [NTU] suggested we should still attempt to get the fastest CPU we can for it. Fair comment. These dudes, http://www.altech.com.au/pages/p52_retail.htm have P200MMXes for $106 and P233MMXes for $109. I think maybe we should invest. More RAM would be nice, but can come later, but IF anyone on committee has any shred of altruism left, or could take time out from feathering their own nests, please buy a new drive for mermaid, so we can keep mako running. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From mtearle at tartarus.uwa.edu.au Thu Sep 2 15:12:49 1999 From: mtearle at tartarus.uwa.edu.au (Mark Tearle) Date: Wed Oct 27 01:17:18 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> Message-ID: On Thu, 2 Sep 1999, David Manchester wrote: > Hmm. > Mtearle thinks its time for a reshuffle - think about how the network > should be laid out WRT the switch & pass on your ideas. > Yes, entropy has caught up with the machine room again. - Rebuilding the racks, maybe a third party rack for *.rcpt.to and ziggy? - Preparations for sensible power (tm) - some network cabling - placement of tape drives, etc Yours Mark -- Mark Tearle - mtearle@tartarus.uwa.edu.au Captain Bipto: We did win, didn't we? Blaznee: No, but if we think fast enough we might just live to lie about it. From fryers at ucc.gu.uwa.edu.au Thu Sep 2 15:25:17 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:18 2004 Subject: [tech] Switches and hubs... In-Reply-To: ; from Mark Tearle on Thu, Sep 02, 1999 at 03:12:49PM +0800 References: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902152517.J14778@ucc.gu.uwa.edu.au> Bingle > A while ago Mark Tearle tapped: > Yes, entropy has caught up with the machine room again. The machne room will always have a high degree of entropy. If someone can show me a machine room that lacks entropy - anywhere - then I will be impressed. > - Rebuilding the racks, maybe a third party rack for *.rcpt.to and ziggy? This is always done. Maybe arrange it so that PC's are not stacked on top of eachother. This will make repairing and upgrading them a lot easier. > - Preparations for sensible power (tm) This will become an eventuality some day - but the people installing it are likely to need a lot of access to the machien room. I forsee much uphevall and downtime when the power is finally installed. I would also imagine that the machine room will be reoptimised several times between now and then. > - some network cabling An excellent idea. > - placement of tape drives, etc Similarly excellent. See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From nick at ucc.gu.uwa.edu.au Thu Sep 2 15:35:25 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:19 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Thu, Sep 02, 1999 at 02:58:50PM +0800 References: <199909020657.OAA05413@typhaon.ucs.uwa.edu.au> <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902153524.X17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 02:58:50PM +0800, David Manchester wrote: > I've got to admit that I missed it. Hence me asking. > Got a pointer to it somewhere..? [...] ----- Well, it could be the router, then the best option would probably be to re-arrange the network to put the router on the 100 network and then router the whole 101 network through it, and reconfigure all our backbone systems of course. My preferred network layout: 100.x incoming | 10M 10M +-------------- router ----------------+ | 100M | 100M | 10/100 switch 10/100 hub delni/hub/... | | | | | | | | | | |||| ||||||||||||| fast boxen and fast but insecure dec net/user net/... one or more boxen, xterms 10Mbps hubs etc, maybe 10M hub maybe with another 10M interface out of the router that I can't draw yet. The 2 x 100Mbps network cards for the router would need committee to approve though, and it would also be nice to run the router in a system that can take 4 x PCI network cards (can the present one?) - the 2 existing 10M ones and 2 new 100M ones (hence it would have to stay FreeBSD for the existing 10M 3com cards). ----- The new 100BaseT network cards might be about $114 for 3Coms, and $60-$70 for D-Link DFE-530TX'es or Netgear tulip's like Grahame's. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From nick at ucc.gu.uwa.edu.au Thu Sep 2 15:44:06 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:19 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902152517.J14778@ucc.gu.uwa.edu.au>; from Simon Fryer on Thu, Sep 02, 1999 at 03:25:17PM +0800 References: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> <19990902152517.J14778@ucc.gu.uwa.edu.au> Message-ID: <19990902154405.Y17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 03:25:17PM +0800, Simon Fryer wrote: [...] > > - Preparations for sensible power (tm) > > This will become an eventuality some day - but the people installing it are > likely to need a lot of access to the machien room. I forsee much uphevall > and downtime when the power is finally installed. I would also imagine > that the machine room will be reoptimised several times between now and then. [...] Well, there's a lot we could do towards it simply by Powerboard Overkill (tm). Neatly attach as many power sockets to each rack as there is space for power using devices. Add a couple for luck. Add a few more in sensible spots on the walls or hanging from the roof. Use the appropriate circuits to power each one and possibly do some actual measurements with a clamp meter. - Placement of the monitor switch and serial terminal(s) Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From fryers at ucc.gu.uwa.edu.au Thu Sep 2 16:13:14 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:19 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902153524.X17478@mermaid.ucc.gu.uwa.edu.au>; from Nick Bannon on Thu, Sep 02, 1999 at 03:35:25PM +0800 References: <199909020657.OAA05413@typhaon.ucs.uwa.edu.au> <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au> <19990902153524.X17478@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902161314.B20618@ucc.gu.uwa.edu.au> Bingle > A while ago Nick Bannon tapped: > Well, it could be the router, then the best option would probably be > to re-arrange the network to put the router on the 100 network and then > router the whole 101 network through it, and reconfigure all our backbone > systems of course. > > My preferred network layout: > > 100.x incoming > | 10M 10M > +-------------- router ----------------+ > | 100M | 100M | > 10/100 switch 10/100 hub delni/hub/... > | | | | | | | | | | |||| ||||||||||||| > fast boxen and fast but insecure dec net/user net/... > one or more boxen, xterms > 10Mbps hubs etc, maybe 10M hub This is good. I can't say I have any problems with this. > maybe with another 10M interface out of the router that I can't draw yet. > > The 2 x 100Mbps network cards for the router would need committee > to approve though, and it would also be nice to run the router in a > system that can take 4 x PCI network cards (can the present one?) - > the 2 existing 10M ones and 2 new 100M ones (hence it would have to stay > FreeBSD for the existing 10M 3com cards). I think the UCC should get a couple of decient 100Mb/s cards. These are something thta UCC will need more and more of in the future. How many exec have not gone to Melbourne for the weekend??? Can Bers, Luyer, Nick and Sam/[TAM] be all in the same place for a few minutes to approve this purchase? > The new 100BaseT network cards might be about $114 for 3Coms, and $60-$70 > for D-Link DFE-530TX'es or Netgear tulip's like Grahame's. Are the Netgear tulips based around a Compaq chipset as their name suggest? If so I can not see any problems with getting these cards. See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From fryers at ucc.gu.uwa.edu.au Thu Sep 2 16:19:55 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:20 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902154405.Y17478@mermaid.ucc.gu.uwa.edu.au>; from Nick Bannon on Thu, Sep 02, 1999 at 03:44:06PM +0800 References: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> <19990902152517.J14778@ucc.gu.uwa.edu.au> <19990902154405.Y17478@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902161955.C20618@ucc.gu.uwa.edu.au> > A while ago Nick Bannon tapped: > On Thu, Sep 02, 1999 at 03:25:17PM +0800, Simon Fryer wrote: > [...] > > > - Preparations for sensible power (tm) > > > > This will become an eventuality some day - but the people installing it are > > likely to need a lot of access to the machien room. I forsee much uphevall > > and downtime when the power is finally installed. I would also imagine > > that the machine room will be reoptimised several times between now and then. > [...] > > Well, there's a lot we could do towards it simply by Powerboard Overkill > (tm). Neatly attach as many power sockets to each rack as there is > space for power using devices. Add a couple for luck. Add a few more in > sensible spots on the walls or hanging from the roof. Use the appropriate > circuits to power each one and possibly do some actual measurements with > a clamp meter. And the Powerboard Overkill is different from current UCC setup in which way exactly! :-) About the clamp meter thingy... I think I put the power cable suitable for a clamp meter on starfish. Has anyone actually measured the power consuption yet? > - Placement of the monitor switch and serial terminal(s) And the serial console switch (I think that is what is it!) - Securing of erwin to something solid so it is no longer in the machine room. - replacement of the network card in Manta - reracking merman and debugging the serial problems with the new CPU card. See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From nick at ucc.gu.uwa.edu.au Thu Sep 2 16:46:38 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:20 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902161955.C20618@ucc.gu.uwa.edu.au>; from Simon Fryer on Thu, Sep 02, 1999 at 04:19:55PM +0800 References: <199909020652.OAA22040@mermaid.ucc.gu.uwa.edu.au> <19990902152517.J14778@ucc.gu.uwa.edu.au> <19990902154405.Y17478@mermaid.ucc.gu.uwa.edu.au> <19990902161955.C20618@ucc.gu.uwa.edu.au> Message-ID: <19990902164638.Z17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 04:19:55PM +0800, Simon Fryer wrote: [...] > And the Powerboard Overkill is different from current UCC setup in which way > exactly! :-) Well.. we've probably already got Too Many powerboards, but still very few actual free sockets. When things get shuffled it's usually hard to find a suitable point and yes we have ended up with power cables straddling the racks, for example, which is really... sub-optimal. If you have plenty of free sockets everywhere you might want a machine, you can avoid some of the uglier "temporary" solutions. > About the clamp meter thingy... I think I put the power cable suitable for > a clamp meter on starfish. Has anyone actually measured the power consuption > yet? Doubt it. [...] > - Securing of erwin to something solid so it is no longer in the machine > room. Ah, that's important. Ian's got some of the details for that but basically we were going to get a quote for the cables from CS and anywhere else we're interested in, and start attaching things to benches and so on. This'll take a bit of money. > - replacement of the network card in Manta > > - reracking merman and debugging the serial problems with the new CPU card. Yes, these are long overdue. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From luyer at ucs.uwa.edu.au Thu Sep 2 16:49:58 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:20 2004 Subject: [tech] Switches and hubs... In-Reply-To: Your message of "Thu, 02 Sep 1999 16:13:14 +0800." <19990902161314.B20618@ucc.gu.uwa.edu.au> Message-ID: <199909020849.QAA07176@typhaon.ucs.uwa.edu.au> fryers wrote: > I think the UCC should get a couple of decient 100Mb/s cards. These are > something thta UCC will need more and more of in the future. How many exec have > not gone to Melbourne for the weekend??? Can Bers, Luyer, Nick and Sam/[TAM] > be all in the same place for a few minutes to approve this purchase? I'm leaving midnight next Tuesday to Melbourne, with my belief of the current balance I believe it is quite reasonable and sensible to get a couple of 3c905s. There will be a committee meeting next Tuesday I hope - Bers should be there, I will be there, and we have to hope for two other committee members to be able to make it. Unfortunately we can't just call a committee meeting straight off, and purchasing decisions can't be made outside a committee meeting, we need 4 days notice to call a committee meeting (our regularly scheduled ones are considered to already have notice given at the meeting before). This does slow down/discourage some purchases and is something I personally would like to see changed - either a provision for special committee meetings with shorter notice or a way for a significant majority of the committee to make purchasing decisions up to some limit without waiting for a committee meeting. David. From mustang at ucc.gu.uwa.edu.au Thu Sep 2 16:55:00 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:21 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902164638.Z17478@mermaid.ucc.gu.uwa.edu.au> from Nick Bannon at "Sep 2, 99 04:46:38 pm" Message-ID: <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au> > On Thu, Sep 02, 1999 at 04:19:55PM +0800, Simon Fryer wrote: > > - replacement of the network card in Manta > > - reracking merman and debugging the serial problems with the new CPU card. > Yes, these are long overdue. > Nick. What succks about Manta's NIC..? I'm sure Prep can donate one. When we want to play Merman = 4/330, I'll be happy to assist. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From gbowland at ucc.gu.uwa.edu.au Thu Sep 2 16:58:10 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:21 2004 Subject: [tech] Switches and hubs... In-Reply-To: <19990902161314.B20618@ucc.gu.uwa.edu.au> References: <199909020657.OAA05413@typhaon.ucs.uwa.edu.au> <199909020658.OAA22057@mermaid.ucc.gu.uwa.edu.au> <19990902153524.X17478@mermaid.ucc.gu.uwa.edu.au> <19990902161314.B20618@ucc.gu.uwa.edu.au> Message-ID: <19990902165809.A21247@mussel.ucc.gu.uwa.edu.au> > On Sep 02, Simon Fryer scrawled : > Are the Netgear tulips based around a Compaq chipset as their name suggest? If > so I can not see any problems with getting these cards. I will have a look at the manual when I get home (if it came with one, I bought it with a PC.) I remember being surprised that it was a Tulip - the card isn't actually called a NetGear Tulip. *waves* -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From nick at ucc.gu.uwa.edu.au Thu Sep 2 17:00:19 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:21 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <199909020711.PAA22105@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Thu, Sep 02, 1999 at 03:11:35PM +0800 References: <199909020711.PAA22105@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902170019.A17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 03:11:35PM +0800, David Manchester wrote: > OK, so I got a faster CPU for it. > [NTU] suggested we should still attempt to get the fastest CPU we can for > it. Fair comment. > These dudes, http://www.altech.com.au/pages/p52_retail.htm have P200MMXes > for $106 and P233MMXes for $109. [...] Hmmm... It's not a _bad_ price, and it's still a desirable thing, but I get the feeling we're in danger of splitting our efforts a bit too much. Death by a thousand cuts. What RAM's in mako? 2x8MB 72pin SIMMs? 1x16? Even a single 32MB SIMM should go down nicely. Maxing it out to 64 would be reasonable. Perhaps we should draw up a wishlist of retail items to buy, total it, prioritise and sanity check it, then buy it in a clump. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From nick at ucc.gu.uwa.edu.au Thu Sep 2 17:03:49 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:21 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Thu, Sep 02, 1999 at 04:55:00PM +0800 References: <19990902164638.Z17478@mermaid.ucc.gu.uwa.edu.au> <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902170347.B17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 04:55:00PM +0800, David Manchester wrote: > What succks about Manta's NIC..? > I'm sure Prep can donate one. Packet loss, particularly on large packets (over a few hundred bytes). We've got two potential replacements from the mib MicroVAX II pile. > When we want to play Merman = 4/330, I'll be happy to assist. Great. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From luyer at ucs.uwa.edu.au Thu Sep 2 17:06:58 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:22 2004 Subject: [tech] Switches and hubs... In-Reply-To: Your message of "Thu, 02 Sep 1999 16:55:00 +0800." <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au> Message-ID: <199909020906.RAA07505@typhaon.ucs.uwa.edu.au> Another thing we should do is either a 68hc11 or something simpler for Starfish's boot-up sequence (pressing space at the relevant point). I've been looking at playing with 12c50[89]'s, the things people make PSX modchips out from. There are some examples easily found on the web using them to do random stuff with rs232, and some simple programmer designs for them (but then, I can't even remember which pin is which on a transistor anymore, so I might drop the idea...). If I end up deciding to build a simple programmer for them it might be an easy option if nobody ends up doing the hc11 thing (hoping that it is up to being connected inline and passing most of the signals through, I haven't really looked into this). David. From mustang at ucc.gu.uwa.edu.au Thu Sep 2 17:09:48 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:22 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <19990902170019.A17478@mermaid.ucc.gu.uwa.edu.au> from Nick Bannon at "Sep 2, 99 05:00:19 pm" Message-ID: <199909020909.RAA22740@mermaid.ucc.gu.uwa.edu.au> > Hmmm... It's not a _bad_ price, and it's still a desirable thing, but I > get the feeling we're in danger of splitting our efforts a bit too much. > Death by a thousand cuts. Yes - that's certainly non-optimal, but as the christians and curious cynical bible readers amongst us will affirm, ask and ye shal receive. If lots of people ask for lots of stuff, odds are some will happen. > What RAM's in mako? 2x8MB 72pin SIMMs? 1x16? Even a single 32MB SIMM > should go down nicely. Maxing it out to 64 would be reasonable. I'd bet 2*8s - finding a fast page 32MB will be tres difficult. Recall its a 486 AFAIK, so EDO's probably out. > Perhaps we should draw up a wishlist of retail items to buy, total it, > prioritise and sanity check it, then buy it in a clump. I concur. My version of this list: 1) HDD for Moray. That's right, moray. It still whinges. Its -the- most important machine in the club. Maybe put the SCSI controller and the 1GB into Moray. 2) HDD for Mako. Ben wants his back, Ben wrote FMS. Be nice to Ben. 3) RAM for Moray. 16MB is tight. 4) SCSI cables. It'd be nice to be able to use the 8mm drives Nick and I `found' for the UCC. 5) 100MB/s NICs. Cheap DEC Tulip cards like the Netgears and Dlinks would be excellent. Just don't use then in Alphas. (no SROM makes DU unhappy) ... I'll let someone more inspired continue. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From fryers at ucc.gu.uwa.edu.au Thu Sep 2 17:11:36 1999 From: fryers at ucc.gu.uwa.edu.au (Simon Fryer) Date: Wed Oct 27 01:17:22 2004 Subject: [tech] Switches and hubs... In-Reply-To: <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Thu, Sep 02, 1999 at 04:55:00PM +0800 References: <19990902164638.Z17478@mermaid.ucc.gu.uwa.edu.au> <199909020855.QAA22671@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990902171136.F20618@ucc.gu.uwa.edu.au> Bingle > A while ago David Manchester tapped: > > On Thu, Sep 02, 1999 at 04:19:55PM +0800, Simon Fryer wrote: > > > > - replacement of the network card in Manta > > > - reracking merman and debugging the serial problems with the new CPU card. > > Yes, these are long overdue. > > Nick. > > What succks about Manta's NIC..? Intermittent traffic action under heavy network activity... > I'm sure Prep can donate one. We have a couple sitting on the shelf in the machine room. It is more a metter of me being there, in the right mood and not being asked to do xxx in ordre to intall/replace them. > When we want to play Merman = 4/330, I'll be happy to assist. This could be a bit of a problem. From what I have seen with it there was no volts action through serial ports A and B. Port A being needed for a console. Surprise surprise ports A and B share a common line driver IC. SO, I nedded the whole lot in a rack so I could work on it and make sure that the driver IC is actually being fed +/- 12V (A trivial check) before obtaining a replacement chip. There are a couple of other bits and pieces that need to be looked at as well with this box like the recomended board layout from sun being totally different to every other board layout for a VME rackmount sun that I have ever seen (ok, I havn't seen that many). I would appreciate the assistance though! See Ya Simon -- ------------------------------------------------------------------------ "Well, an engineer is not concerned with the truth; that is left to philosophers and theologians: the prime concern of an engineer is the utility of the final product." Lectures on the Electrical Properties of Materials, L.Solymar, D.Walsh From luyer at ucs.uwa.edu.au Thu Sep 2 17:16:57 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:22 2004 Subject: [tech] Mermaid re-animated In-Reply-To: Your message of "Thu, 02 Sep 1999 17:09:48 +0800." <199909020909.RAA22740@mermaid.ucc.gu.uwa.edu.au> Message-ID: <199909020916.RAA07830@typhaon.ucs.uwa.edu.au> Another comment about memory for mako - if we're talking about buying it, even if it will accept EDO - it is damn expensive. eg - it was cheaper to get a Celeron 400, motherboard (AT form factor), slot card and 128Mb PC100 DIMM ($520 inc tax) than to get 128Mb of EDO DRAM ($580 inc tax for 4x32Mb at the tame time). The EDO was "cost price" from the supplier who usually gives us the best memory prices. People with older stock still have slightly lower prices on EDO, but not too much. David. From alastair at ucc.gu.uwa.edu.au Thu Sep 2 19:12:30 1999 From: alastair at ucc.gu.uwa.edu.au (Alastair Irvine) Date: Wed Oct 27 01:17:22 2004 Subject: [tech] Duplicated tech list postings Message-ID: <19990902191230.G20681@ucc.gu.uwa.edu.au> Has anyone noticed than some posts to the tech list (without apparent rhyme or reason) get distributed twice? The only thing different about the pairs of messages are their topmost Received header fields, e.g.: Received: from moray (daemon@localhost [127.0.0.1]) by ucc.gu.uwa.edu.au (8.9.3/8.9.3/Debian/GNU) with ESMTP id RAA06256; Thu, 2 Sep 1999 17:12:06 +0800 and Received: from moray (mailman@localhost [127.0.0.1]) by ucc.gu.uwa.edu.au (8.9.3/8.9.3/Debian/GNU) with ESMTP id RAA06268; Thu, 2 Sep 1999 17:12:15 +0800 These were from Simon Fryer's followup message(s) to the "[tech] Switches and hubs...." thread, regarding Manta dodginess. Same with two of mustang's posts. This has got me thinking... Could this whole problem be caused by /home/other/mailman/scripts/post ? PS -- The tech archives can be accessed at /home/other/mailman/archives/private/tech . The abovementioned messages are not duplicated here. When I clicked on the archive link within the tech list web page, I got sent to an error page saying that shouldn't have sent me to and that webmaster had been notified. Did anybody on the webmaster alias actually receive this alleged e-mail? -- ... "Any place with an all-night, drive-through taxidermist has got to be weird" -- Billy Connolly, about LA _______________________________________________________________________ | | | -=*Alastair Irvine*=- WWW page: http://members.xoom.com/alastair/ | | System Administrator Magic:tG player, coder, RPGer, net-nut | | e-mail: alastair@ucc.gu.uwa.edu.au or alastair.irvine@usa.net | |_______________________________________________________________________| From ben at ucc.gu.uwa.edu.au Thu Sep 2 20:08:05 1999 From: ben at ucc.gu.uwa.edu.au (Ben Rampling) Date: Wed Oct 27 01:17:23 2004 Subject: [tech] Mermaid re-animated References: <199909020916.RAA07830@typhaon.ucs.uwa.edu.au> Message-ID: <004901bef53b$d09e3f60$fa6b08cb@WSB> > Another comment about memory for mako - if we're talking about buying it, even > if it will accept EDO - it is damn expensive. > > eg - it was cheaper to get a Celeron 400, motherboard (AT form factor), > slot card and 128Mb PC100 DIMM ($520 inc tax) than to get 128Mb of EDO DRAM > ($580 inc tax for 4x32Mb at the tame time). The EDO was "cost price" from > the supplier who usually gives us the best memory prices. People with older > stock still have slightly lower prices on EDO, but not too much. Good point - why buy old expensive upgrades when we could buy new ones that'll be useful for much longer. If Mako becames a C400/128, we'd solve the problem of mailing lists and Moray's performance problems, Mako's current performance problems, we'd be able to move to PostgreSQL, and we'd have some of Mako's old hardware that could be used for something less demanding. If Moray has the slots available, we could bump it up to 32Mb with Mako's two 8Mb's. If we want to look at buying a motherboard, 128Mb of RAM and the Celeron 400 in the next couple of weeks, I'm quite happy to donate the $200 or so that the club would need to buy one of the 7200rpm Quantum drives previously mentioned. Regards, Ben Rampling From ben at ucc.gu.uwa.edu.au Thu Sep 2 20:31:44 1999 From: ben at ucc.gu.uwa.edu.au (Ben Rampling) Date: Wed Oct 27 01:17:23 2004 Subject: [tech] Hi Message-ID: <006301bef53f$1e91d0d0$fa6b08cb@WSB> Testing, Testing, 123. :) From japester at ucc.gu.uwa.edu.au Fri Sep 3 09:56:44 1999 From: japester at ucc.gu.uwa.edu.au (Jean-Paul Blaquiere) Date: Wed Oct 27 01:17:23 2004 Subject: Host key on *beowulf* [was: Re: [tech] Host key on mussel] In-Reply-To: <19990902142129.F20681@ucc.gu.uwa.edu.au> References: <19990902103504.A20681@ucc.gu.uwa.edu.au> <19990902104202.A9746@wyvern.com.au> <19990902111438.V17478@mermaid.ucc.gu.uwa.edu.au> <19990902142129.F20681@ucc.gu.uwa.edu.au> Message-ID: <19990903095643.A7517@beowulf.ucc.gu.uwa.edu.au> > On Sep 02, Alastair Irvine scrawled : > Warning: Remote host denied X11 forwarding, perhaps xauth program could not be run on the server side. > hmmm I must look at this then. > > As a general rule - even during upgrades you can avoid breaking the host > > key. I don't know whether this is strictly desirable or not. > > Hmm... I guess JP didn't think about that. No problem. > not one generally can recover when the filesystem is trashed, requiring a *full* reinstall from scratch. apolgies for any problems occured. *wave* -- Jean-Paul Blaquiere japester@ucc.gu.uwa.edu.au gotta www.wibble.org again From nick at ucc.gu.uwa.edu.au Fri Sep 3 11:10:59 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:23 2004 Subject: [tech] Duplicated tech list postings In-Reply-To: <19990902191230.G20681@ucc.gu.uwa.edu.au>; from Alastair Irvine on Thu, Sep 02, 1999 at 07:12:30PM +0800 References: <19990902191230.G20681@ucc.gu.uwa.edu.au> Message-ID: <19990903111058.C17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 07:12:30PM +0800, Alastair Irvine wrote: > Has anyone noticed than some posts to the tech list (without apparent rhyme > or reason) get distributed twice? The only thing different about the pairs > of messages are their topmost Received header fields, e.g.: Yes... It happens on the other lists too. Don't know how or why, I'm afraid. [...] > This has got me thinking... Could this whole problem be caused by > /home/other/mailman/scripts/post ? Quite possibly, but as it doesn't happen every time, I imagine it's some combination of factors. > PS -- The tech archives can be accessed at > /home/other/mailman/archives/private/tech . The abovementioned > messages are not duplicated here. [...] Yeah - there is a message to webmasters, but the simple answer is that it isn't archived (properly?) yet. This should be easy enough to fix, but no-one's looked at it yet. Not a hugely helpful email, but no, it's not just you. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From mustang at ucc.gu.uwa.edu.au Fri Sep 3 11:12:54 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:23 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <004901bef53b$d09e3f60$fa6b08cb@WSB> from Ben Rampling at "Sep 2, 99 08:08:05 pm" Message-ID: <199909030312.LAA25827@mermaid.ucc.gu.uwa.edu.au> > Good point - why buy old expensive upgrades when we could buy new ones > that'll be useful for much longer. I'm all for it. > If Mako becames a C400/128, we'd solve > the problem of mailing lists and Moray's performance problems, Mako's > current performance problems, we'd be able to move to PostgreSQL, and we'd > have some of Mako's old hardware that could be used for something less > demanding. If Moray has the slots available, we could bump it up to 32Mb > with Mako's two 8Mb's. Actually, I'd advocate stealing Moray's RAM (as there are no more slots free) and installing that into Mako and doing a brain transplant. > If we want to look at buying a motherboard, 128Mb of RAM and the Celeron 400 > in the next couple of weeks, I'm quite happy to donate the $200 or so that > the club would need to buy one of the 7200rpm Quantum drives previously > mentioned. Urm, read the above sentence again, recall what spurned this whole conversation & see if you're as puzzled as I. I think I understand Ben's rationale :) Anyway - C400A/RAM/bigfoot=Mako, P150/64MB/6GB disk=Mermaid, Mako+RAM+SCSI=moray... how's that sound..? Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From nick at ucc.gu.uwa.edu.au Fri Sep 3 11:50:56 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:24 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <199909020909.RAA22740@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Thu, Sep 02, 1999 at 05:09:48PM +0800 References: <19990902170019.A17478@mermaid.ucc.gu.uwa.edu.au> <199909020909.RAA22740@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990903115055.D17478@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 02, 1999 at 05:09:48PM +0800, David Manchester wrote: [...] > My version of this list: > > 1) HDD for Moray. That's right, moray. It still whinges. Its -the- most > important machine in the club. > Maybe put the SCSI controller and the 1GB into Moray. > > 2) HDD for Mako. Ben wants his back, Ben wrote FMS. Be nice to Ben. > > 3) RAM for Moray. 16MB is tight. > > 4) SCSI cables. It'd be nice to be able to use the 8mm drives Nick and I > `found' for the UCC. > > 5) 100MB/s NICs. Cheap DEC Tulip cards like the Netgears and Dlinks would > be excellent. Just don't use then in Alphas. (no SROM makes DU unhappy) 6) 4 port dumb serial card for moray so as to centralise the evil dispense serial stuff. I might have a suitable one. This might be better achieved by other means. David Manchester wrote: > Ben Rampling wrote: > > If we want to look at buying a motherboard, 128Mb of RAM and the Celeron 400 > > in the next couple of weeks, I'm quite happy to donate the $200 or so that > > the club would need to buy one of the 7200rpm Quantum drives previously > > mentioned. > > Urm, read the above sentence again, recall what spurned this whole > conversation & see if you're as puzzled as I. > I think I understand Ben's rationale :) > > Anyway - C400A/RAM/bigfoot=Mako, P150/64MB/6GB disk=Mermaid, > +Mako+RAM+SCSI=moray... how's that sound..? 7) This, or some modification of this. I'd just like to make another plea for naming scheme sanity - by the time a machine gets a new MB, CPU, RAM, I/O controller and HDD it probably should have a new name... ::-/ If it happens one by one over years, that's one thing, but still... Hopefully a webmaster will get around to updating ; http://www.ucc.gu.uwa.edu.au/machines/ sometime. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From mustang at ucc.gu.uwa.edu.au Fri Sep 3 13:05:38 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:24 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <19990903115055.D17478@mermaid.ucc.gu.uwa.edu.au> from Nick Bannon at "Sep 3, 99 11:50:56 am" Message-ID: <199909030505.NAA26229@mermaid.ucc.gu.uwa.edu.au> > On Thu, Sep 02, 1999 at 05:09:48PM +0800, David Manchester wrote: > > 1) HDD for Moray. That's right, moray. It still whinges. Its -the- most > > 2) HDD for Mako. Ben wants his back, Ben wrote FMS. Be nice to Ben. > > 3) RAM for Moray. 16MB is tight. > > 4) SCSI cables. It'd be nice to be able to use the 8mm drives Nick and I > > 5) 100MB/s NICs. Cheap DEC Tulip cards like the Netgears and Dlinks would > 6) 4 port dumb serial card for moray so as to centralise the evil dispense > I'd just like to make another plea for naming scheme sanity - by the time > a machine gets a new MB, CPU, RAM, I/O controller and HDD it probably > should have a new name... ::-/ If it happens one by one over years, > that's one thing, but still... Similarly though - names have become indicative of roles. i.e. Moray's a trusted services box. Marlin's a backup services & general user machine. IF we build a new moray, its still moray. Moray and Marlin the PC have had at least 4 motherboards between them since I've been about the club.. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From ben at ucc.gu.uwa.edu.au Fri Sep 3 14:17:45 1999 From: ben at ucc.gu.uwa.edu.au (Ben Rampling) Date: Wed Oct 27 01:17:24 2004 Subject: [ucc] Re: [tech] Mermaid re-animated In-Reply-To: <199909030600.OAA19748@typhaon.ucs.uwa.edu.au> from David Luyer at "Sep 3, 99 02:00:12 pm" Message-ID: <199909030617.OAA26562@mermaid.ucc.gu.uwa.edu.au> > > The data for fms can't be run accross nfs, afaik, and the source tends to > > be on both the ucc cvs mirror and my own. > > It could if you were using Linux 2.2 on Mako. Linux 2.2 can magically optimise a database accessing its data volume over NFS? I'm astounded. Let's switch now! Regards, Ben Rampling From luyer at ucs.uwa.edu.au Fri Sep 3 14:22:25 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:24 2004 Subject: [ucc] Re: [tech] Mermaid re-animated In-Reply-To: Your message of "Fri, 03 Sep 1999 14:17:45 +0800." <199909030617.OAA26562@mermaid.ucc.gu.uwa.edu.au> Message-ID: <199909030622.OAA20044@typhaon.ucs.uwa.edu.au> > > > The data for fms can't be run accross nfs, afaik, and the source tends to > > > be on both the ucc cvs mirror and my own. > > > > It could if you were using Linux 2.2 on Mako. > > Linux 2.2 can magically optimise a database accessing its data volume over > NFS? I'm astounded. Let's switch now! You said it _can't_ be run over NFS. The only possible technical reason something can't be run over NFS is if it absolutely requires locking operations or raw device access. I'm pretty sure you're not using raw device access, hence if it _can't_ be run over NFS it would be a locking issue. Running Linux 2.2 kernels and knfs everywhere would fix that :-) David. PS: don't trim the smiley in the follow up next time... From nick at ucc.gu.uwa.edu.au Fri Sep 3 15:26:18 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:24 2004 Subject: [tech] Mermaid re-animated In-Reply-To: <199909030505.NAA26229@mermaid.ucc.gu.uwa.edu.au>; from David Manchester on Fri, Sep 03, 1999 at 01:05:38PM +0800 References: <19990903115055.D17478@mermaid.ucc.gu.uwa.edu.au> <199909030505.NAA26229@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990903152617.F17478@mermaid.ucc.gu.uwa.edu.au> On Fri, Sep 03, 1999 at 01:05:38PM +0800, David Manchester wrote: [...] > Similarly though - names have become indicative of roles. > i.e. Moray's a trusted services box. Marlin's a backup services & general > user machine. > > IF we build a new moray, its still moray. Well, that's another way of looking at it, but as far as I'm concerned that's what CNAMEs are for - we've got about 50 of the things... moray's coke, ftp and flame (now an extra A record). mako's www, cvs and fms (A record). mola's home, nfs and loghost. > Moray and Marlin the PC have had at least 4 motherboards between them > since I've been about the club.. Well, exactly, that's the point. Marlin was even a SPARC at one stage! When I think of marlin, I think of a 5MB 486DX-33, with a mono VGA monitor, running 386BSD and Linux 0.99.3. Perhaps a copy of xmosaic causing it to swap hideously. Moray's upgrades have been more spaced. Some memory here, a motherboard there. It only took the jump to 486DX4-100 from a 486DX-33 relatively recently. It probably still deserves its name, though I wouldn't swear to it. Ah well. Whatever. If you _do_ do major surgery to an existing machine, it'd be nice if you updated the machine's web page. If it's a brand new machine, it's OK to wait until it's stabilised, I guess. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From luyer at ucs.uwa.edu.au Fri Sep 3 16:32:49 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:24 2004 Subject: [tech] More old stuff: DELNI, DEMPR and DECServer 300's Message-ID: <199909030832.QAA22683@typhaon.ucs.uwa.edu.au> I've got a DELNI, DEMPR and a few DECServer 300's for UCC via EE Workshops from the airport. The coolest thing about them being that they have CAA asset tags :-) Anyway, someone please come and pick them up from the UCS office. David. From mustang at ucc.gu.uwa.edu.au Sun Sep 5 11:31:06 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:25 2004 Subject: [tech] Re: [wheel] 2nd HDD on erwin In-Reply-To: <003901bef6d3$fb83be00$f400a8c0@jedi> from Anil Sharma at "Sep 4, 99 08:49:50 pm" Message-ID: <199909050331.LAA01157@mermaid.ucc.gu.uwa.edu.au> > This is now exported as erwin:/opt and is mounted on azure as /opt. > I have installed netscape on it and will soon try and migrate window > managers to it. > Of course the real fun starts when I try and make these work from /opt for > all users. Didn't /opt already exist on Erwin, i.e. its a default directory. I hope you moved everything out of it onto the 2GB ..? Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From maset at ucc.gu.uwa.edu.au Sun Sep 5 21:48:29 1999 From: maset at ucc.gu.uwa.edu.au (Anil Sharma) Date: Wed Oct 27 01:17:25 2004 Subject: [tech] Re: [wheel] 2nd HDD on erwin In-Reply-To: <199909050331.LAA01157@mermaid.ucc.gu.uwa.edu.au> Message-ID: [DAVE *grin*] > Didn't /opt already exist on Erwin, i.e. its a default directory. > I hope you moved everything out of it onto the 2GB ..? > Oh forgot to mention. Tarballs of original /opt for erwin and azure are sitting in my home directory conveniently called erwinopt.tar and azureopt.tar. Maset the Grandiose. ----------------------------------------------- Without suffering, how can one appreciate happiness? And how would we mark the depths of our sorrow, without the light of hope? From mustang at ucc.gu.uwa.edu.au Sun Sep 5 22:34:52 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:25 2004 Subject: [tech] Re: [wheel] 2nd HDD on erwin In-Reply-To: from Anil Sharma at "Sep 5, 99 09:48:29 pm" Message-ID: <199909051434.WAA04044@mermaid.ucc.gu.uwa.edu.au> > [DAVE *grin*] > > Didn't /opt already exist on Erwin, i.e. its a default directory. > > I hope you moved everything out of it onto the 2GB ..? > > > Oh forgot to mention. Tarballs of original /opt for erwin and azure are > sitting in my home directory conveniently called erwinopt.tar and > azureopt.tar. OK, fly. Maybe check some of the permissions/ownership on that stuff, though. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From gbowland at ucc.gu.uwa.edu.au Mon Sep 6 11:07:04 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:25 2004 Subject: [tech] Monitor on the Be Message-ID: <19990906110704.A13408@mussel.ucc.gu.uwa.edu.au> The monitor on the Be is dying, slowly and painfully. Turning it on leads to crackling sounds and noticable interference on the speakers and other monitors around the room. Its off for now until its verified that its even safe... See you, -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From gbowland at ucc.gu.uwa.edu.au Mon Sep 6 13:30:01 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:25 2004 Subject: [tech] Disk space on mermaid Message-ID: <19990906133000.A7257@mermaid.ucc.gu.uwa.edu.au> Mermaid was down to almost no free space - in the order of 15Mb. Thats fairly alarmingly low, so I've removed something large - the JDK. We've got 45Mb free now so unless someone puts huge files in /tmp we'll be okay :) Anyone wanting to do Java development should probably use mussel - using Java on mermaid tended to slag it. *waves* -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From nick at ucc.gu.uwa.edu.au Mon Sep 6 13:37:57 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] Disk space on mermaid In-Reply-To: <19990906133000.A7257@mermaid.ucc.gu.uwa.edu.au>; from Grahame Bowland on Mon, Sep 06, 1999 at 01:30:01PM +0800 References: <19990906133000.A7257@mermaid.ucc.gu.uwa.edu.au> Message-ID: <19990906133756.U17478@mermaid.ucc.gu.uwa.edu.au> On Mon, Sep 06, 1999 at 01:30:01PM +0800, Grahame Bowland wrote: > Mermaid was down to almost no free space - in the order of 15Mb. Thats fairly > alarmingly low, so I've removed something large - the JDK. We've got 45Mb free > now so unless someone puts huge files in /tmp we'll be okay :) There's also a 53MB /usr/local/jdk1.2 - that's probably next on the chopping block. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From dichro-69db471 at eris.rcpt.to Mon Sep 6 13:41:43 1999 From: dichro-69db471 at eris.rcpt.to (Mikolaj J. Habryn) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] Disk space on mermaid In-Reply-To: Nick Bannon's message of "Mon, 6 Sep 1999 13:37:57 +0800" References: <19990906133000.A7257@mermaid.ucc.gu.uwa.edu.au> <19990906133756.U17478@mermaid.ucc.gu.uwa.edu.au> Message-ID: >>>>> "NB" == Nick Bannon writes: NB> There's also a 53MB /usr/local/jdk1.2 - that's probably next NB> on the chopping block. I'd suggest clobbering the emacs stuff - that should clear out quite a bit more, and running emacs on mermaid is really an act of desperation. m. From japester at ucc.gu.uwa.edu.au Tue Sep 7 12:18:42 1999 From: japester at ucc.gu.uwa.edu.au (Jean-Paul Blaquiere) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] socks Message-ID: <19990907121842.A13127@beowulf.ucc.gu.uwa.edu.au> i am trying to socksify beowulf, but I hav no idea what to do about startup scripts an config files. Could someone please enlighten me? *wave* - Jean-Paul Blaquiere japester@ucc.gu.uwa.edu.au gotta www.wibble.org again From mustang at ucc.gu.uwa.edu.au_no.spam.please_ Wed Sep 15 13:06:21 1999 From: mustang at ucc.gu.uwa.edu.au_no.spam.please_ (mustang@ucc.gu.uwa.edu.au_no.spam.please_) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] (fwd) DPC Super Buy: RICOH 7040A (20/4/4) CD Writer $369 Message-ID: <199909150506.NAA24504@mermaid.ucc.gu.uwa.edu.au> -- forwarded message -- Path: news.uwa.edu.au!news.mel.connect.com.au!news.mel.aone.net.au!newsfeed.aone.net.au!ozemail.com.au!not-for-mail From: "Dkmaj" Newsgroups: aus.ads.forsale.computers.new Subject: DPC Super Buy: RICOH 7040A (20/4/4) CD Writer $369 Lines: 10 X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.00.2615.200 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200 Message-ID: <3LFD3.14830$1E2.102328@ozemail.com.au> NNTP-Posting-Host: slora2p29.ozemail.com.au X-Trace: ozemail.com.au 937371711 203.108.146.77 (Wed, 15 Sep 1999 15:01:51 EST) NNTP-Posting-Date: Wed, 15 Sep 1999 15:01:51 EST Organization: OzEmail Ltd, Australia Distribution: world Date: Wed, 15 Sep 1999 15:01:10 +1000 Xref: news.uwa.edu.au aus.ads.forsale.computers.new:23098 DPC can offer this special buy, full retail pack, with Adaptec Easy CD Creator included. Low price for this week of $369. CD Rewriter 20X (read) 4X (write)4X (rewrite) Email: DPC@Tech-center.com Web: http://www.treeoflifemusic.org.au/dpc/dpchome.htm -- end of forwarded message -- -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From gbowland at ucc.gu.uwa.edu.au Wed Sep 15 13:42:25 1999 From: gbowland at ucc.gu.uwa.edu.au (Grahame Bowland) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] CD-R prices Message-ID: <19990915134225.B1607@mussel.ucc.gu.uwa.edu.au> Here they are: -------------------------------------------------------------------------- RTV Computers (rtv.iinet.net.au): * CRW4416 4x4x16 Internal SCSI CD-RW Drive OEM Bare Drive - $499 Austin (austin-computers.com.au) * * Plextor CDR 8x20, 4MB Buffer SCSI $690 DPC (http://www.treeoflifemusic.org.au/dpc) * Trinix (trinix.com.au) * Panasonic CW7502 SCSI 4W/8R Drive only $443 (tax-ex $371) * Teac 6W/24R SCSI Internal Drive $567 (tax-ex $482) * Teac 8W/24R SCSI Internal Drive Only $780 (tax-ex $661) * Teac 4W/24R SCSI Internal Drive Only New Low Price $413 (tax-ex $351) --------------------------------------------------------------------- This is all copied and pasted from websites. If any information is missing, blame their webmasters! -- Grahame Bowland /---------------------------------------------------------------------\ | Email: gbowland@ucc.gu.uwa.edu.au gbowland@iee.org.uk | | gbowland@tartarus.uwa.edu.au gbowland@hotmail.com | | Web: http://users.wantree.com.au/~bowest/gmb.html | \---------------------------------------------------------------------/ From yakk at yakk.net.au Thu Sep 16 13:45:52 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:26 2004 Subject: [tech] flame, moray's kernel and transparent proxying Message-ID: <19990916134552.A16719@yakk.net.au> Hi, I've worked out a way to give flame its own IP number without needing to give it its own machine. We can use the transparent proxying code in linux to redirect all requests to port 23 with a destination of flame's IP# (I grabbed 130.95.13.10 - it doesn't seem to be used) to port 4242. We can do this with all the other services that flame runs. I've set up ip aliasing on moray (its now using 130.95.13.(9|10)), but moray's kernel doesn't seem to have the nececary firewall options enabled to do the port redirection. Where/how was moray's kernel compiled? moray:/usr/src/linux seems to be 2.0.30, but moray is running 2.0.37... Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From warrick at wyvern.com.au Thu Sep 16 13:47:18 1999 From: warrick at wyvern.com.au (Warrick Mitchell) Date: Wed Oct 27 01:17:27 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916134552.A16719@yakk.net.au> References: <19990916134552.A16719@yakk.net.au> Message-ID: <19990916134718.A4680@wyvern.com.au> On Thu, Sep 16, 1999 at 01:45:52PM +0800, Ian McKellar wrote: > Hi, > > I've worked out a way to give flame its own IP number without needing to > give it its own machine. We can use the transparent proxying code in linux > to redirect all requests to port 23 with a destination of flame's IP# (I > grabbed 130.95.13.10 - it doesn't seem to be used) to port 4242. We can do > this with all the other services that flame runs. I've set up ip aliasing > on moray (its now using 130.95.13.(9|10)), but moray's kernel doesn't seem > to have the nececary firewall options enabled to do the port redirection. [root@quigon]:~# nslookup 130.95.13.10 Server: asia14-ofpe.perth.oilfield.slb.com Address: 163.184.30.2 Name: q.rcpt.to Address: 130.95.13.10 It was in use and you knew about it unless someone has switched the machine off. I would suggest you choose another ip address. Seeya, Warrick From yakk at yakk.net.au Thu Sep 16 13:57:23 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:27 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916134718.A4680@wyvern.com.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> Message-ID: <19990916135723.B16719@yakk.net.au> On Thu, Sep 16, 1999 at 01:47:18PM +0800, Warrick Mitchell wrote: > > [root@quigon]:~# nslookup 130.95.13.10 > Server: asia14-ofpe.perth.oilfield.slb.com > Address: 163.184.30.2 > > Name: q.rcpt.to > Address: 130.95.13.10 > > It was in use and you knew about it unless someone has switched the machine off. > > I would suggest you choose another ip address. Ahh - well, I just looked though the UCC zone file... I'll try to think up another one. Whats q up to these days? Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From nick at ucc.gu.uwa.edu.au Thu Sep 16 14:02:04 1999 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:17:27 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916134552.A16719@yakk.net.au>; from Ian McKellar on Thu, Sep 16, 1999 at 01:45:52PM +0800 References: <19990916134552.A16719@yakk.net.au> Message-ID: <19990916140204.K21708@mermaid.ucc.gu.uwa.edu.au> On Thu, Sep 16, 1999 at 01:45:52PM +0800, Ian McKellar wrote: [...] > Where/how was moray's kernel compiled? moray:/usr/src/linux seems to be > 2.0.30, but moray is running 2.0.37... Compiling moray's own kernel on moray is probably unneccessarily slow and space consuming. If you're in a kernel compiling mood, use mussel and 2.0.38. You probably need to use gcc272 rather than the default egcs. Nick. -- Nick Bannon | "I made this letter longer than usual because nick@it.net.au | I lack the time to make it shorter." - Pascal From dunc at rcpt.to Thu Sep 16 14:02:48 1999 From: dunc at rcpt.to (Duncan Sargeant) Date: Wed Oct 27 01:17:27 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916134718.A4680@wyvern.com.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> Message-ID: <19990916140247.A563@mussel.ucc.gu.uwa.edu.au> Warrick Mitchell wrote on Thu September 16, at 13:47 +0800: > On Thu, Sep 16, 1999 at 01:45:52PM +0800, Ian McKellar wrote: > > Hi, > > > > I've worked out a way to give flame its own IP number without needing to > > give it its own machine. We can use the transparent proxying code in linux > > to redirect all requests to port 23 with a destination of flame's IP# (I > > grabbed 130.95.13.10 - it doesn't seem to be used) to port 4242. We can do > > this with all the other services that flame runs. I've set up ip aliasing > > on moray (its now using 130.95.13.(9|10)), but moray's kernel doesn't seem > > to have the nececary firewall options enabled to do the port redirection. Can you do this without ditzy kernel options, with rlinetd dichro? > [root@quigon]:~# nslookup 130.95.13.10 > Server: asia14-ofpe.perth.oilfield.slb.com > Address: 163.184.30.2 > > Name: q.rcpt.to Address: 130.95.13.10 > > It was in use and you knew about it unless someone has switched the > machine off. > > I would suggest you choose another ip address. It was also marlin a while back. ah well :-) ,dunc From warrick at wyvern.com.au Thu Sep 16 13:59:34 1999 From: warrick at wyvern.com.au (Warrick Mitchell) Date: Wed Oct 27 01:17:27 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916135723.B16719@yakk.net.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916135723.B16719@yakk.net.au> Message-ID: <19990916135934.A4721@wyvern.com.au> On Thu, Sep 16, 1999 at 01:57:23PM +0800, Ian McKellar wrote: > On Thu, Sep 16, 1999 at 01:47:18PM +0800, Warrick Mitchell wrote: > > > > [root@quigon]:~# nslookup 130.95.13.10 > > Server: asia14-ofpe.perth.oilfield.slb.com > > Address: 163.184.30.2 > > > > Name: q.rcpt.to > > Address: 130.95.13.10 > > > > It was in use and you knew about it unless someone has switched the machine off. > > > > I would suggest you choose another ip address. > > Ahh - well, I just looked though the UCC zone file... "Assumption is the mother of all fuck up's" did you ping it? :) checking the reverse dns file is good practice :) > I'll try to think up another one. Whats q up to these days? q.rcpt.to is a mail server and shell box for rcpt.to thats in progress. Seeya, Warrick From yakk at yakk.net.au Thu Sep 16 14:14:18 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:28 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916135934.A4721@wyvern.com.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916135723.B16719@yakk.net.au> <19990916135934.A4721@wyvern.com.au> Message-ID: <19990916141418.C16719@yakk.net.au> On Thu, Sep 16, 1999 at 01:59:34PM +0800, Warrick Mitchell wrote: > > > > "Assumption is the mother of all fuck up's" did you ping it? :) checking the reverse dns file is good practice :) I pinged it. It didn't pong. Its funny - I can ping it from outside UCC, but not from moray. Its on the right segment isn't it? > > > > > I'll try to think up another one. Whats q up to these days? > > q.rcpt.to is a mail server and shell box for rcpt.to thats in progress. Hmm - any chance of hosting an rcpt.to mail/dns server outside of the UCC for reliability/cost reduction? Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From warrick at wyvern.com.au Thu Sep 16 14:13:16 1999 From: warrick at wyvern.com.au (Warrick Mitchell) Date: Wed Oct 27 01:17:28 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916141418.C16719@yakk.net.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916135723.B16719@yakk.net.au> <19990916135934.A4721@wyvern.com.au> <19990916141418.C16719@yakk.net.au> Message-ID: <19990916141316.A4824@wyvern.com.au> On Thu, Sep 16, 1999 at 02:14:18PM +0800, Ian McKellar wrote: > On Thu, Sep 16, 1999 at 01:59:34PM +0800, Warrick Mitchell wrote: > > > > > > > > "Assumption is the mother of all fuck up's" did you ping it? :) checking the reverse dns file is good practice :) > > I pinged it. It didn't pong. Its funny - I can ping it from outside UCC, but > not from moray. Its on the right segment isn't it? Hmm it is pinging correctly now that eth0:0 is down and not pretending to be 130.95.13.10 q.rcpt.to can ping moray. moray can ping q.rcpt.to Seeya, Warrick From yakk at yakk.net.au Thu Sep 16 14:27:15 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:28 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916140247.A563@mussel.ucc.gu.uwa.edu.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916140247.A563@mussel.ucc.gu.uwa.edu.au> Message-ID: <19990916142715.D16719@yakk.net.au> On Thu, Sep 16, 1999 at 02:02:48PM +0800, Duncan Sargeant wrote: > Warrick Mitchell wrote on Thu September 16, at 13:47 +0800: > > On Thu, Sep 16, 1999 at 01:45:52PM +0800, Ian McKellar wrote: > > > Hi, > > > > > > I've worked out a way to give flame its own IP number without needing to > > > give it its own machine. We can use the transparent proxying code in linux > > > to redirect all requests to port 23 with a destination of flame's IP# (I > > > grabbed 130.95.13.10 - it doesn't seem to be used) to port 4242. We can do > > > this with all the other services that flame runs. I've set up ip aliasing > > > on moray (its now using 130.95.13.(9|10)), but moray's kernel doesn't seem > > > to have the nececary firewall options enabled to do the port redirection. > > Can you do this without ditzy kernel options, with rlinetd dichro? No. We could set up a tunnel, but that would mean that flame wouldn't know the real IP that the person came from. Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From dunc at rcpt.to Thu Sep 16 14:37:40 1999 From: dunc at rcpt.to (Duncan Sargeant) Date: Wed Oct 27 01:17:28 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916142715.D16719@yakk.net.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916140247.A563@mussel.ucc.gu.uwa.edu.au> <19990916142715.D16719@yakk.net.au> Message-ID: <19990916143740.B563@mussel.ucc.gu.uwa.edu.au> Ian McKellar wrote on Thu September 16, at 14:27 +0800: > On Thu, Sep 16, 1999 at 02:02:48PM +0800, Duncan Sargeant wrote: > > Warrick Mitchell wrote on Thu September 16, at 13:47 +0800: > > > On Thu, Sep 16, 1999 at 01:45:52PM +0800, Ian McKellar wrote: > > > > Hi, > > > > > > > > I've worked out a way to give flame its own IP number without > > > > needing to give it its own machine. We can use the > > > > transparent proxying code in linux to redirect all requests to > > > > port 23 with a destination of flame's IP# (I grabbed > > > > 130.95.13.10 - it doesn't seem to be used) to port 4242. We > > > > can do this with all the other services that flame runs. I've > > > > set up ip aliasing on moray (its now using 130.95.13.(9|10)), > > > > but moray's kernel doesn't seem to have the nececary firewall > > > > options enabled to do the port redirection. > > > > Can you do this without ditzy kernel options, with rlinetd dichro? > > No. We could set up a tunnel, but that would mean that flame > wouldn't know the real IP that the person came from. Redirection is a hack for cracks; you use redirection to get around firewalls. It would be much cleaner if you just bound the services to their correct ports on /flame/'s address. rlinetd lets you do this with the interfaces command. ,dunc From luyer at ucs.uwa.edu.au Thu Sep 16 14:52:30 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:28 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 14:37:40 +0800." <19990916143740.B563@mussel.ucc.gu.uwa.edu.au> Message-ID: <199909160652.OAA27661@typhaon.ucs.uwa.edu.au> > It would be much cleaner if you just bound the services to their > correct ports on /flame/'s address. rlinetd lets you do this with the > interfaces command. You could easily modify the driver and moray's inetd to bind() things correctly to support flame.ucc and moray.ucc on the one machine. David. From yakk at yakk.net.au Thu Sep 16 14:59:18 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:29 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916143740.B563@mussel.ucc.gu.uwa.edu.au> References: <19990916134552.A16719@yakk.net.au> <19990916134718.A4680@wyvern.com.au> <19990916140247.A563@mussel.ucc.gu.uwa.edu.au> <19990916142715.D16719@yakk.net.au> <19990916143740.B563@mussel.ucc.gu.uwa.edu.au> Message-ID: <19990916145918.A17078@yakk.net.au> On Thu, Sep 16, 1999 at 02:37:40PM +0800, Duncan Sargeant wrote: > > Redirection is a hack for cracks; you use redirection to get around > firewalls. > > It would be much cleaner if you just bound the services to their > correct ports on /flame/'s address. rlinetd lets you do this with the > interfaces command. > It may be a hack, but its a more elegant hack than hacking MudOS to bind to a particlar ip, and: * running MudOS as root or * hacking the kernel to allow non-root to bind to low ports Running MudOS as root would be an unacceptable security risk, and I feel that enabling a kernel compile time option is less of a hack than writing a kernel source hack to violate one of the most basic premises of the unix security model. It would allow someone who got clever enough to get access to the socket stuff in flame have effectively superuser access to stuff like dispense (which trusts connections based on source port). Doing it at the firewall level is the most elegant way I've come up with (and I've been thinking about this on and off for a couple of years). Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From yakk at yakk.net.au Thu Sep 16 14:59:56 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:29 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <199909160652.OAA27661@typhaon.ucs.uwa.edu.au> References: <19990916143740.B563@mussel.ucc.gu.uwa.edu.au> <199909160652.OAA27661@typhaon.ucs.uwa.edu.au> Message-ID: <19990916145956.B17078@yakk.net.au> On Thu, Sep 16, 1999 at 02:52:30PM +0800, David Luyer wrote: > > It would be much cleaner if you just bound the services to their > > correct ports on /flame/'s address. rlinetd lets you do this with the > > interfaces command. > > You could easily modify the driver and moray's inetd to bind() things > correctly to support flame.ucc and moray.ucc on the one machine. > yes, but port 23 and port 80? Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From mustang at ucc.gu.uwa.edu.au Thu Sep 16 14:55:42 1999 From: mustang at ucc.gu.uwa.edu.au (David Manchester) Date: Wed Oct 27 01:17:29 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <199909160652.OAA27661@typhaon.ucs.uwa.edu.au> from David Luyer at "Sep 16, 1999 02:52:30 pm" Message-ID: <199909160655.OAA04568@mussel.ucc.gu.uwa.edu.au> > > It would be much cleaner if you just bound the services to their > > correct ports on /flame/'s address. rlinetd lets you do this with the > > interfaces command. > > You could easily modify the driver and moray's inetd to bind() things > correctly to support flame.ucc and moray.ucc on the one machine. > Or we could migrate moray and flame onto separate boxes & we could get the UCC's IPs back from rcpt.to Moray is underpowered as is. Cheers /dave -- / David Manchester [TDH] Netware/UNIX droid. \ Tell someone who's interested, tell someone who can keep their lunch digested Tell someone who wants your conversation, tell someone who doesn't regard you \as an argument for compulsory sterilisation." [TISM], `How Do I Love Thee?'/ From luyer at ucs.uwa.edu.au Thu Sep 16 15:00:53 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:29 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 14:59:56 +0800." <19990916145956.B17078@yakk.net.au> Message-ID: <199909160700.PAA28104@typhaon.ucs.uwa.edu.au> > On Thu, Sep 16, 1999 at 02:52:30PM +0800, David Luyer wrote: > > > It would be much cleaner if you just bound the services to their > > > correct ports on /flame/'s address. rlinetd lets you do this with the > > > interfaces command. > > > > You could easily modify the driver and moray's inetd to bind() things > > correctly to support flame.ucc and moray.ucc on the one machine. > > > yes, but port 23 and port 80? Make moray's inetd and httpd bind port 80 to specific interfaces. Probably sendmail too if you want to do real SMTP on flame. Make flames driver bind to specific interfaces port 80, 4242 and 23 (and others). Multiple incoming socket support isn't hard if it isn't there already. David. From yakk at yakk.net.au Thu Sep 16 15:07:40 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:29 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <199909160700.PAA28104@typhaon.ucs.uwa.edu.au> References: <19990916145956.B17078@yakk.net.au> <199909160700.PAA28104@typhaon.ucs.uwa.edu.au> Message-ID: <19990916150740.C17078@yakk.net.au> On Thu, Sep 16, 1999 at 03:00:53PM +0800, David Luyer wrote: > > Make flames driver bind to specific interfaces port 80, 4242 and 23 (and > others). Multiple incoming socket support isn't hard if it isn't there > already. > Can the flame driver do this running as user 'flame'? Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From luyer at ucs.uwa.edu.au Thu Sep 16 15:04:46 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:30 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 14:59:18 +0800." <19990916145918.A17078@yakk.net.au> Message-ID: <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> > It may be a hack, but its a more elegant hack than hacking MudOS to bind to > a particlar ip, and: > * running MudOS as root > or > * hacking the kernel to allow non-root to bind to low ports or * using a launcher app (the standard way, see innd and many others) fd0 = bind(flame_ip:23); fd1 = bind(flame_ip:80); exec(flame, fd0#, fd1#); David. From dichro-fc44bc08 at eris.rcpt.to Thu Sep 16 15:08:57 1999 From: dichro-fc44bc08 at eris.rcpt.to (Mikolaj J. Habryn) Date: Wed Oct 27 01:17:30 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: David Luyer's message of "Thu, 16 Sep 1999 15:04:46 +0800" References: <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> Message-ID: >>>>> "DL" == David Luyer writes: DL> fd0 = bind(flame_ip:23); fd1 = bind(flame_ip:80); exec(flame, DL> fd0#, fd1#); *blink* That's some seriously weird-ass libc bindings you're using... :) m. From yakk at yakk.net.au Thu Sep 16 15:14:40 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:30 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> References: <19990916145918.A17078@yakk.net.au> <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> Message-ID: <19990916151440.D17078@yakk.net.au> On Thu, Sep 16, 1999 at 03:04:46PM +0800, David Luyer wrote: > > It may be a hack, but its a more elegant hack than hacking MudOS to bind to > > a particlar ip, and: > > * running MudOS as root > > or > > * hacking the kernel to allow non-root to bind to low ports > or > * using a launcher app (the standard way, see innd and many others) > > fd0 = bind(flame_ip:23); > fd1 = bind(flame_ip:80); > exec(flame, fd0#, fd1#); > Ahh, cool, but is this less of a hack than using the transparent proxy code? Also, with the transparent proxy code we can add new services without rebooting, recompiling flame, or even restarting the driver :) Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From luyer at ucs.uwa.edu.au Thu Sep 16 15:23:48 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:30 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 15:14:40 +0800." <19990916151440.D17078@yakk.net.au> Message-ID: <199909160723.PAA28730@typhaon.ucs.uwa.edu.au> > Ahh, cool, but is this less of a hack than using the transparent proxy > code? Also, with the transparent proxy code we can add new services without > rebooting, recompiling flame, or even restarting the driver :) No, but it can be implemented without a kernel compile and reboot :-) David. From dunc at rcpt.to Thu Sep 16 15:28:34 1999 From: dunc at rcpt.to (Duncan Sargeant) Date: Wed Oct 27 01:17:30 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916151440.D17078@yakk.net.au> References: <19990916145918.A17078@yakk.net.au> <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> <19990916151440.D17078@yakk.net.au> Message-ID: <19990916152834.C563@mussel.ucc.gu.uwa.edu.au> Ian McKellar wrote on Thu September 16, at 15:14 +0800: > On Thu, Sep 16, 1999 at 03:04:46PM +0800, David Luyer wrote: > > > It may be a hack, but its a more elegant hack than hacking MudOS to bind to > > > a particlar ip, and: > > > * running MudOS as root > > > or > > > * hacking the kernel to allow non-root to bind to low ports > > or > > * using a launcher app (the standard way, see innd and many others) > > > > fd0 = bind(flame_ip:23); > > fd1 = bind(flame_ip:80); > > exec(flame, fd0#, fd1#); > > Ahh, cool, but is this less of a hack than using the transparent proxy > code? Yes. > Also, with the transparent proxy code we can add new services without > rebooting, recompiling flame, or even restarting the driver :) *boggle* How were you planning to enable the transparent proxy code without recompiling the kernel or rebooting moray? ,dunc From luyer at ucs.uwa.edu.au Thu Sep 16 15:33:03 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:31 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 15:28:34 +0800." <19990916152834.C563@mussel.ucc.gu.uwa.edu.au> Message-ID: <199909160733.PAA28925@typhaon.ucs.uwa.edu.au> I wrote: > > > fd0 = bind(flame_ip:23); > > > fd1 = bind(flame_ip:80); > > > exec(flame, fd0#, fd1#); The challenge, of course, is to get this integrated into mainline MudOS. See if you can make an elegant patch which makes it so that if the driver is called as driver -p 3:80 -p 4:23 [other options] then it stores somewhere the information that FD 3 is bound to port 80 and FD 4 is bound to port 23 and then when someone tries to do a bind to said ports it switches file descriptors underneath them. David. From yakk at yakk.net.au Thu Sep 16 15:58:29 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:31 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <19990916152834.C563@mussel.ucc.gu.uwa.edu.au> References: <19990916145918.A17078@yakk.net.au> <199909160704.PAA28335@typhaon.ucs.uwa.edu.au> <19990916151440.D17078@yakk.net.au> <19990916152834.C563@mussel.ucc.gu.uwa.edu.au> Message-ID: <19990916155829.E17078@yakk.net.au> On Thu, Sep 16, 1999 at 03:28:34PM +0800, Duncan Sargeant wrote: > > > Also, with the transparent proxy code we can add new services without > > rebooting, recompiling flame, or even restarting the driver :) > > *boggle* How were you planning to enable the transparent proxy code > without recompiling the kernel or rebooting moray? > No, once its running, to add additional services (the flame FNORDd, or whatever) wouldn't require extensive breaking. Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From luyer at ucs.uwa.edu.au Thu Sep 16 16:23:03 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:31 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: Your message of "Thu, 16 Sep 1999 15:58:29 +0800." <19990916155829.E17078@yakk.net.au> Message-ID: <199909160823.QAA29874@typhaon.ucs.uwa.edu.au> > No, once its running, to add additional services (the flame FNORDd, or > whatever) wouldn't require extensive breaking. How often are you going to want a new service on a secure port? David. From yakk at yakk.net.au Thu Sep 16 16:36:45 1999 From: yakk at yakk.net.au (Ian McKellar) Date: Wed Oct 27 01:17:31 2004 Subject: [tech] flame, moray's kernel and transparent proxying In-Reply-To: <199909160823.QAA29874@typhaon.ucs.uwa.edu.au> References: <19990916155829.E17078@yakk.net.au> <199909160823.QAA29874@typhaon.ucs.uwa.edu.au> Message-ID: <19990916163645.G17078@yakk.net.au> On Thu, Sep 16, 1999 at 04:23:03PM +0800, David Luyer wrote: > > No, once its running, to add additional services (the flame FNORDd, or > > whatever) wouldn't require extensive breaking. > > How often are you going to want a new service on a secure port? > Well, I do about every 10th time I log into flame :) Ian -- Ian McKellar | Email: yakk(a)yakk.net | Web: http://www.yakk.net/ Prefix: +61 8 | Fax/VoiceMail: 9265 0821 | Home: 9389 9162 | Work: 9380 3688 From luyer at ucs.uwa.edu.au Fri Sep 17 23:10:08 1999 From: luyer at ucs.uwa.edu.au (David Luyer) Date: Wed Oct 27 01:17:31 2004 Subject: [tech] switch & 100Mbit hub... Message-ID: <199909171510.XAA15214@typhaon.ucs.uwa.edu.au> the switch is now happily switching 7 x 10Mbit ports plus mussel; * the coax thingy * the 12-port hub which is now only the incoming connection on coax * mussel (@100Mbits :-) * mola * kraken * q * enki * uccrouter These computers were chosen by being the only 10 UTP-connected things on the UCC backbone. The 100Mbit hub now contains the 2 SGIs, UCCrouter and the other hub. (moving towards the diagram I posted out before) David.