From lathiat at sixlabs.org Fri Aug 1 18:13:35 2003 From: lathiat at sixlabs.org (Trent Lloyd) Date: Wed Oct 27 01:27:41 2004 Subject: [tech] DVD media Message-ID: <20030801101335.GA23199@sixlabs.org> Heyas, I can get DVD-R for $2.50 in packs of 10 Anyone want any? I only want about 5 so if someone wants the other 5 (or UCC does?) then i can grab them and you can give me $12.50 -- Cheers, Trent Sixlabs GPG Fingerprint: 6EFF FE20 BFE3 F721 6F26 B13A F974 3E56 04AB 3C5D I accept GPG encrypted mail, do you? o_O From maset at ucc.gu.uwa.edu.au Tue Aug 5 07:01:12 2003 From: maset at ucc.gu.uwa.edu.au (Anil Sharma) Date: Wed Oct 27 01:27:43 2004 Subject: [tech] WAIX connectivity Message-ID: <5.2.0.9.0.20030805070017.0272ae18@mooneye.ucc.gu.uwa.edu.au> Is there something wrong with the *.waix.ucc addresses? I don't seem to be able to connect to them from arachnet. Cheers, Anil. BSc. (Neuroscience) Hons. From trs80 at ucc.gu.uwa.edu.au Tue Aug 5 11:24:28 2003 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Wed Oct 27 01:27:43 2004 Subject: [tech] WAIX connectivity In-Reply-To: <5.2.0.9.0.20030805070017.0272ae18@mooneye.ucc.gu.uwa.edu.au> Message-ID: On Tue, 5 Aug 2003, Anil Sharma wrote: > Is there something wrong with the *.waix.ucc addresses? I don't seem to be > able to connect to them from arachnet. Yes, there is. Aarnet has done something to block them, Grahame is investigating. -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC President http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 / From adrian at ucc.gu.uwa.edu.au Tue Aug 5 11:40:37 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:44 2004 Subject: [tech] WAIX connectivity In-Reply-To: References: <5.2.0.9.0.20030805070017.0272ae18@mooneye.ucc.gu.uwa.edu.au> Message-ID: <20030805034037.GA402545@morwong.ucc.gu.uwa.edu.au> On Tue, Aug 05, 2003, James Andrewartha wrote: > On Tue, 5 Aug 2003, Anil Sharma wrote: > > > Is there something wrong with the *.waix.ucc addresses? I don't seem to be > > able to connect to them from arachnet. > > Yes, there is. Aarnet has done something to block them, Grahame is > investigating. Don't hold your breath. If you need to use the *.waix.ucc addresses to get into UCC Let me know and I'll try to setup a bounce on a WAIX machine. adrian From grahame at angrygoats.net Tue Aug 12 16:15:45 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:44 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <20030812062030.GA30189@ucc.gu.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> Message-ID: <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> On Tue, 2003-08-12 at 14:20, Davyd wrote: > Here they are, shiny happy minutes holding hands... I've been advised by AARNET that they are unwilling to have the 172.26.32/19 addresses crossing administrative boundaries. They are trying to build a resilient PARNET core, with two POPs, two links to AARNET and two links to WAIX. The 172 addresses are a bit of a pain for them. If someone can find a location in WAIX willing to run a Cisco style IP-IP tunnel, I'll set that up. Please don't ask WAIX (eg Gav) to do this, as I don't think it should be their problem. Cheers Grahame billy:~# cat /var/log/haiku*|wc -l 114262 From adrian at ucc.gu.uwa.edu.au Tue Aug 12 16:21:27 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:44 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> Message-ID: <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> On Tue, Aug 12, 2003, Grahame Bowland wrote: > On Tue, 2003-08-12 at 14:20, Davyd wrote: > > Here they are, shiny happy minutes holding hands... > > I've been advised by AARNET that they are unwilling to have the > 172.26.32/19 addresses crossing administrative boundaries. They are > trying to build a resilient PARNET core, with two POPs, two links to > AARNET and two links to WAIX. The 172 addresses are a bit of a pain for > them. > > If someone can find a location in WAIX willing to run a Cisco style > IP-IP tunnel, I'll set that up. Please don't ask WAIX (eg Gav) to do > this, as I don't think it should be their problem. Sure thing. Grab me about it after work, I have a crisco speakething the Beegeepee. Adrian From nick at ucc.gu.uwa.edu.au Tue Aug 12 16:46:49 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:44 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> On Tue, Aug 12, 2003 at 04:21:27PM +0800, Adrian Chadd wrote: > Sure thing. Grab me about it after work, I have a crisco speakething > the Beegeepee. There's another option: 1. Tunnel 172.26.42.0/24 into WAIX 2. Use WAIX BGP data to unfirewall 130.95.13.0/24 <-> WAIX on hydra or its replacement. Which do you think is the better option? Can option 2 go horribly wrong and cost us transit traffic, or even if it can, can we automagically notice that and shut it down? Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From grahame at angrygoats.net Tue Aug 12 23:20:23 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:45 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> Message-ID: <1060701623.4695.2.camel@typhaon.ucs.uwa.edu.au> On Tue, 2003-08-12 at 16:46, Nick Bannon wrote: > On Tue, Aug 12, 2003 at 04:21:27PM +0800, Adrian Chadd wrote: > > Sure thing. Grab me about it after work, I have a crisco speakething > > the Beegeepee. > > There's another option: > 1. Tunnel 172.26.42.0/24 into WAIX > 2. Use WAIX BGP data to unfirewall 130.95.13.0/24 <-> WAIX on hydra or > its replacement. > > Which do you think is the better option? Can option 2 go horribly wrong > and cost us transit traffic, or even if it can, can we automagically > notice that and shut it down? 1. Was my original idea, IPIP tunnel to something in WAIX and get them to advertise it for us. It's not the best, but it would work 2. I don't know if putting a HUGE access list onto hydra is a good idea, at least not until we get a replacement. Then again, someone might as well right the code. I can get you a feed of what WAIX is from the the UWA perspective easily enough, say with two days notice. Two would be cute to implement and a lot more flexible, but also a lot more work. From grahame at angrygoats.net Tue Aug 12 23:21:28 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:45 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <1060701623.4695.2.camel@typhaon.ucs.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> <1060701623.4695.2.camel@typhaon.ucs.uwa.edu.au> Message-ID: <1060701688.4695.5.camel@typhaon.ucs.uwa.edu.au> On Tue, 2003-08-12 at 23:20, Grahame Bowland wrote: > On Tue, 2003-08-12 at 16:46, Nick Bannon wrote: > > On Tue, Aug 12, 2003 at 04:21:27PM +0800, Adrian Chadd wrote: > > > Sure thing. Grab me about it after work, I have a crisco speakething > > > the Beegeepee. > > > > There's another option: > > 1. Tunnel 172.26.42.0/24 into WAIX > > 2. Use WAIX BGP data to unfirewall 130.95.13.0/24 <-> WAIX on hydra or > > its replacement. > > > > Which do you think is the better option? Can option 2 go horribly wrong > > and cost us transit traffic, or even if it can, can we automagically > > notice that and shut it down? > > 1. Was my original idea, IPIP tunnel to something in WAIX and get them > to advertise it for us. It's not the best, but it would work > > 2. I don't know if putting a HUGE access list onto hydra is a good idea, > at least not until we get a replacement. Then again, someone might as > well right the code. I can get you a feed of what WAIX is from the the > UWA perspective easily enough, say with two days notice. BTW I didn't get Adrian's mail that you replied to.. ack. Something dodgy is with the mailing lists. Where did adrian send it? From davyd at zdlcomputing.com Wed Aug 13 09:14:07 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:27:46 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> Message-ID: <1060737247.1326.8.camel@pingu> On Tue, 2003-08-12 at 16:46, Nick Bannon wrote: > There's another option: > 1. Tunnel 172.26.42.0/24 into WAIX > 2. Use WAIX BGP data to unfirewall 130.95.13.0/24 <-> WAIX on hydra or > its replacement. From a "the way it should work" perspective, I prefer (2). It would allow us to trash the .waix.ucc subdomain, and just have machines work from the correct places without multiple domain names (the way it should be). It there a way to guarentee the BGP of WAIX? and to firewall accordingly? Could it be done in the UWA core? or would it be better off done on bananabox (looking to become hydra's decendant), or some new piece of cisco kit. -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030813/212d7042/attachment.pgp From adrian at ucc.gu.uwa.edu.au Wed Aug 13 09:52:26 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:46 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <1060737247.1326.8.camel@pingu> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> <1060737247.1326.8.camel@pingu> Message-ID: <20030813015225.GG108377@morwong.ucc.gu.uwa.edu.au> On Wed, Aug 13, 2003, Davyd wrote: > It there a way to guarentee the BGP of WAIX? and to firewall > accordingly? Could it be done in the UWA core? or would it be better off > done on bananabox (looking to become hydra's decendant), or some new > piece of cisco kit. Nope. its pretty close. I _think_ the aarnet guys do billing based off bgp tables too, so you'd want to make sure you're seeing what they're seeing. which, for the most case, will be right. if you want guaratees, run an ipip tunnel. Adrian From grahame at angrygoats.net Wed Aug 13 15:39:32 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:47 2004 Subject: [tech] Re: [ucc] Committee Minutes 2003-08-12 In-Reply-To: <1060737247.1326.8.camel@pingu> References: <20030812062030.GA30189@ucc.gu.uwa.edu.au> <1060676145.12836.29.camel@typhaon.ucs.uwa.edu.au> <20030812082127.GB108377@morwong.ucc.gu.uwa.edu.au> <20030812084649.GL3981@morwong.ucc.gu.uwa.edu.au> <1060737247.1326.8.camel@pingu> Message-ID: <1060760372.7625.7.camel@solitaire> On Wed, 2003-08-13 at 09:14, Davyd wrote: > On Tue, 2003-08-12 at 16:46, Nick Bannon wrote: > > > There's another option: > > 1. Tunnel 172.26.42.0/24 into WAIX > > 2. Use WAIX BGP data to unfirewall 130.95.13.0/24 <-> WAIX on hydra or > > its replacement. > > From a "the way it should work" perspective, I prefer (2). > It would allow us to trash the .waix.ucc subdomain, and just have > machines work from the correct places without multiple domain names (the > way it should be). > > It there a way to guarentee the BGP of WAIX? and to firewall > accordingly? Could it be done in the UWA core? or would it be better off > done on bananabox (looking to become hydra's decendant), or some new > piece of cisco kit. Fundamentally, you use BGP to determine *your* routing table. This is the decision of where to send *outbound* traffic. However, you are charged on *inbound* traffic. If asymmetric routing happens, you'll get charged for "free" routes that appear in the BGP table of WAIX. It's not that rare for traffic to be sent to WAIX and come back through Optus / AARNET, I think there are a few known offenders. Do the IPIP tunnel, I'll talk to Adrian about getting it set up. -- Grahame Bowland all those things hath mine hand made and all the people say amen praise ye the From trs80 at ucc.gu.uwa.edu.au Wed Aug 13 22:03:28 2003 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Wed Oct 27 01:27:47 2004 Subject: [tech] morwong's disks Message-ID: Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in use, with /home wholly using one of the 9s. This means we have 4x 9gig disks available for /home. There are several options: o Software RAID 5 across all 4 disks. Gives 27gig usable space, some redundancy, some CPU hit. o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives 18gig usable space, good redundancy, little CPU usage (comparatively). o Two RAID 0 sets containing 2 disks each, added to the one domain. Gives 36gig usable space, no redundancy, good performance, little CPU usage. There are two other options similar to the last one, namely one RAID 0 set containing 4 disks (higher throughput, but it'll block more since AdvFS only sees the one logical device), or just 4 disks in the domain (lower throughput, more spindles). So, what do we want to do? One other thing to note is that the first option will require deleting /home from morwong and then restoring it from backup, while the others can be done simply by shuffling volumes around in home_domain. -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC President http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 / From elixxir at ucc.asn.au Wed Aug 13 22:26:39 2003 From: elixxir at ucc.asn.au (Paul Marinceu) Date: Wed Oct 27 01:27:47 2004 Subject: [tech] morwong's disks Message-ID: <1060784799.2755.19.camel@delonge> On Wed, 2003-08-13 at 22:03, James Andrewartha wrote: > Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in > use, with /home wholly using one of the 9s. This means we have 4x 9gig > disks available for /home. There are several options: > > o Software RAID 5 across all 4 disks. Gives 27gig usable space, some > redundancy, some CPU hit. > > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > 18gig usable space, good redundancy, little CPU usage (comparatively). > > o Two RAID 0 sets containing 2 disks each, added to the one domain. Gives > 36gig usable space, no redundancy, good performance, little CPU usage. > > There are two other options similar to the last one, namely one RAID 0 set > containing 4 disks (higher throughput, but it'll block more since AdvFS > only sees the one logical device), or just 4 disks in the domain (lower > throughput, more spindles). > > So, what do we want to do? Well, I'm not much of a raid guru here, but: I say, let's go for all out RAID-0 w/ all 4 disks. (if people think advfs sucks, then make it 2 disks each) What's the use of _some_ redundancy if it's too damn slow to be useable reasonably. (bleh, this post should be in tech) -- Paul Marinceu http://elixxir.ucc.asn.au From nick at ucc.gu.uwa.edu.au Wed Aug 13 22:37:54 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:48 2004 Subject: [tech] morwong's disks In-Reply-To: References: Message-ID: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> On Wed, Aug 13, 2003 at 10:03:28PM +0800, James Andrewartha wrote: > Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in > use, with /home wholly using one of the 9s. This means we have 4x 9gig > disks available for /home. There are several options: Great! We could still do with a SCA SBB to replace that 4 with a 9... > o Software RAID 5 across all 4 disks. Gives 27gig usable space, some > redundancy, some CPU hit. Not just a CPU hit, a significant write performance hit. I wouldn't worry if it was /services, but I don't think that's best for /home. Worse, it binds all those disks into one big clump that it's hard to grow/shrink/fiddle with, without having the same space again spare. > o Two RAID 0 sets containing 2 disks each, added to the one domain. Gives > 36gig usable space, no redundancy, good performance, little CPU usage. Ummm. Well. It's what we've lived with up until now, but you just can't trust disks these days, and these aren't even new disks... I think that leaves us a prudent but high-performance choice of: > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > 18gig usable space, good redundancy, little CPU usage (comparatively). [...] We can also turn mirroring on and off, so we can copy the files from one set to the other and remove any fragmentation. Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From grahame at angrygoats.net Wed Aug 13 22:53:48 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:48 2004 Subject: [tech] morwong's disks In-Reply-To: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> Message-ID: <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> On Wed, 2003-08-13 at 22:37, Nick Bannon wrote: > On Wed, Aug 13, 2003 at 10:03:28PM +0800, James Andrewartha wrote: > > Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in > > use, with /home wholly using one of the 9s. This means we have 4x 9gig > > disks available for /home. There are several options: > > Great! > > We could still do with a SCA SBB to replace that 4 with a 9... > > > o Software RAID 5 across all 4 disks. Gives 27gig usable space, some > > redundancy, some CPU hit. > > Not just a CPU hit, a significant write performance hit. I wouldn't > worry if it was /services, but I don't think that's best for /home. > > Worse, it binds all those disks into one big clump that it's hard to > grow/shrink/fiddle with, without having the same space again spare. > > > o Two RAID 0 sets containing 2 disks each, added to the one domain. Gives > > 36gig usable space, no redundancy, good performance, little CPU usage. > > Ummm. Well. It's what we've lived with up until now, but you just can't > trust disks these days, and these aren't even new disks... We've been damned lucky. Similar Seagate 9Gbs fail all the time at UCS and we have an air conditioned machine room. I'd say go the 2 * RAID-1 disk sets. Seems like the best option. > I think that leaves us a prudent but high-performance choice of: > > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > > 18gig usable space, good redundancy, little CPU usage (comparatively). > [...] > > We can also turn mirroring on and off, so we can copy the files from > one set to the other and remove any fragmentation. You shouldn't need to. If there is free space then the fragmentation should be automatically kept down. From bonfire at bur.st Wed Aug 13 23:01:19 2003 From: bonfire at bur.st (Paul Day) Date: Wed Oct 27 01:27:48 2004 Subject: [tech] morwong's disks In-Reply-To: <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> Message-ID: On Wed, 13 Aug 2003, Grahame Bowland wrote: > > Ummm. Well. It's what we've lived with up until now, but you just can't > > trust disks these days, and these aren't even new disks... > > We've been damned lucky. Similar Seagate 9Gbs fail all the time at UCS > and we have an air conditioned machine room. I'd say go the 2 * RAID-1 > disk sets. Seems like the best option. Yeah, I've had similar problems with those 9GB SCSI Seagates. You can get them pretty cheap from the US though. PD -- Paul Day Web: www.bur.st/~bonfire GPG Key ID: 2EF4ED23 From bonfire at bur.st Wed Aug 13 23:03:41 2003 From: bonfire at bur.st (Paul Day) Date: Wed Oct 27 01:27:48 2004 Subject: [tech] morwong's disks In-Reply-To: References: <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> Message-ID: On Wed, 13 Aug 2003, Paul Day wrote: > > We've been damned lucky. Similar Seagate 9Gbs fail all the time at UCS > > and we have an air conditioned machine room. I'd say go the 2 * RAID-1 > > disk sets. Seems like the best option. > > Yeah, I've had similar problems with those 9GB SCSI Seagates. You can get > them pretty cheap from the US though. Actually, James has just informed me that I actually provided one of them - so for the reocrd, I claim no responsibility if it dies and strongly recommend RAID1. ;) PD -- Paul Day Web: www.bur.st/~bonfire GPG Key ID: 2EF4ED23 From adrian at ucc.gu.uwa.edu.au Thu Aug 14 01:05:14 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:49 2004 Subject: [tech] morwong's disks In-Reply-To: References: Message-ID: <20030813170513.GA8474@morwong.ucc.gu.uwa.edu.au> On Wed, Aug 13, 2003, James Andrewartha wrote: > Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in > use, with /home wholly using one of the 9s. This means we have 4x 9gig > disks available for /home. There are several options: > > o Software RAID 5 across all 4 disks. Gives 27gig usable space, some > redundancy, some CPU hit. > > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > 18gig usable space, good redundancy, little CPU usage (comparatively). I'd do this. If the free space was kept high, life would be all sweet. IO performance wouldn't suck. Since we have many, many small files I'd suggest keeping something rather fat like a 2 and a bit gig freespace allotment. I'm sure you can do that kinda shit under advfs but if you can't would a big group quota suffice? I know wheel users can still abuse their rewt but still.. FYI, this evenings run: Top 5 Homedirs by usage: /home/wheel/ 974615 /home/wheel//nick/ 666941 /home/wheel//trs80/ 353375 /home/wheel//mtearle/ 322122 /home/wheel//matt/ 300982 /home/wheel//yakk/ Top 5 Homedirs by usage: /home/ucc/ 504031 /home/ucc//knight/ 397116 /home/ucc//tieryn/ 374589 /home/ucc//chas/ 370093 /home/ucc//zarquin/ 357053 /home/ucc//packrat/ thats 2.6gig by the wheel users and 2.0 gig by the gumby users. Thats, like, 30% of space being used by 10 users. And they consistent increase nightly. Damned mail. Adrian From grahame at angrygoats.net Thu Aug 14 01:30:03 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:49 2004 Subject: [tech] morwong's disks In-Reply-To: <20030813170513.GA8474@morwong.ucc.gu.uwa.edu.au> References: <20030813170513.GA8474@morwong.ucc.gu.uwa.edu.au> Message-ID: <1060795803.7869.4.camel@solitaire> On Thu, 2003-08-14 at 01:05, Adrian Chadd wrote: > On Wed, Aug 13, 2003, James Andrewartha wrote: > > Morwong now has 5x 9gig and 1x4 gig disks. Currently, 2 9s and one are in > > use, with /home wholly using one of the 9s. This means we have 4x 9gig > > disks available for /home. There are several options: > > > > o Software RAID 5 across all 4 disks. Gives 27gig usable space, some > > redundancy, some CPU hit. > > > > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > > 18gig usable space, good redundancy, little CPU usage (comparatively). > > I'd do this. If the free space was kept high, life would be all sweet. > IO performance wouldn't suck. Since we have many, many small files > I'd suggest keeping something rather fat like a 2 and a bit gig freespace > allotment. > > I'm sure you can do that kinda shit under advfs but if you can't would > a big group quota suffice? I know wheel users can still abuse their > rewt but still.. > > FYI, this evenings run: > > Top 5 Homedirs by usage: /home/wheel/ > 974615 /home/wheel//nick/ > 666941 /home/wheel//trs80/ > 353375 /home/wheel//mtearle/ > 322122 /home/wheel//matt/ > 300982 /home/wheel//yakk/ > > > Top 5 Homedirs by usage: /home/ucc/ > 504031 /home/ucc//knight/ > 397116 /home/ucc//tieryn/ > 374589 /home/ucc//chas/ > 370093 /home/ucc//zarquin/ > 357053 /home/ucc//packrat/ > > thats 2.6gig by the wheel users and 2.0 gig by the gumby users. > Thats, like, 30% of space being used by 10 users. And they consistent > increase nightly. Damned mail. And there's no real need for it - especially if it's just storing large email archives and random gunk. I know it builds up, and if it's source trees for UCC stuff then it makes sense, but... :-) Everyone in wheel has other places to archive mail with more storage than morwong. -- Grahame Bowland broad ways they shall bear the names of his two sons and the greatness of From davidb-7654 at rcpt.to Thu Aug 14 10:47:10 2003 From: davidb-7654 at rcpt.to (David Basden) Date: Wed Oct 27 01:27:49 2004 Subject: [tech] morwong's disks In-Reply-To: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030814024710.GB85987@misato.xware.cx> On Wed, Aug 13, 2003 at 10:37:54PM +0800, Nick Bannon wrote: > I think that leaves us a prudent but high-performance choice of: > > o Two RAID 1 sets containing 2 disks each, added to the one domain. Gives > > 18gig usable space, good redundancy, little CPU usage (comparatively). > [...] This is the option I would rather go with as well. I have had just too many disks die to use anything other than RAID for something backed up as little, and used by so many people as UCC /home. David From nick at ucc.gu.uwa.edu.au Thu Aug 14 11:09:15 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:50 2004 Subject: [tech] morwong's disks In-Reply-To: <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> Message-ID: <20030814030915.GA2952@morwong.ucc.gu.uwa.edu.au> On Wed, Aug 13, 2003 at 10:53:48PM +0800, Grahame Bowland wrote: [...] > > We can also turn mirroring on and off, so we can copy the files from > > one set to the other and remove any fragmentation. > > You shouldn't need to. If there is free space then the fragmentation > should be automatically kept down. That won't affect old fragmentation, though, will it? just stop it getting worse? Quotas (or group quotas - handy, I'd forgotten about those, Adrian!) sound like the way to give reserved root space on AdvFS. I have and will pay for a bigger quota, though. ::-) Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From adrian at ucc.gu.uwa.edu.au Thu Aug 14 11:43:09 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:50 2004 Subject: [tech] morwong's disks In-Reply-To: <20030814030915.GA2952@morwong.ucc.gu.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> <20030814030915.GA2952@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030814034309.GD8474@morwong.ucc.gu.uwa.edu.au> On Thu, Aug 14, 2003, Nick Bannon wrote: > On Wed, Aug 13, 2003 at 10:53:48PM +0800, Grahame Bowland wrote: > [...] > > > We can also turn mirroring on and off, so we can copy the files from > > > one set to the other and remove any fragmentation. > > > > You shouldn't need to. If there is free space then the fragmentation > > should be automatically kept down. > > That won't affect old fragmentation, though, will it? just stop it > getting worse? > > Quotas (or group quotas - handy, I'd forgotten about those, Adrian!) > sound like the way to give reserved root space on AdvFS. I have and > will pay for a bigger quota, though. ::-) Buy a seperate disk, let us mount it as /home/wheel/nick ? Adrian From grahame at angrygoats.net Thu Aug 14 12:57:27 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:50 2004 Subject: [tech] morwong's disks In-Reply-To: <20030814030915.GA2952@morwong.ucc.gu.uwa.edu.au> References: <20030813143753.GA3751@morwong.ucc.gu.uwa.edu.au> <1060786428.1616.3.camel@typhaon.ucs.uwa.edu.au> <20030814030915.GA2952@morwong.ucc.gu.uwa.edu.au> Message-ID: <1060837047.7638.7.camel@typhaon.ucs.uwa.edu.au> On Thu, 2003-08-14 at 11:09, Nick Bannon wrote: > On Wed, Aug 13, 2003 at 10:53:48PM +0800, Grahame Bowland wrote: > [...] > > > We can also turn mirroring on and off, so we can copy the files from > > > one set to the other and remove any fragmentation. > > > > You shouldn't need to. If there is free space then the fragmentation > > should be automatically kept down. > > That won't affect old fragmentation, though, will it? just stop it > getting worse? > > Quotas (or group quotas - handy, I'd forgotten about those, Adrian!) > sound like the way to give reserved root space on AdvFS. I have and > will pay for a bigger quota, though. ::-) Try breaking your mail up; morwong tends to die when you run mutt over your huge inbox. :) From trs80 at ucc.gu.uwa.edu.au Thu Aug 14 21:26:29 2003 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Wed Oct 27 01:27:51 2004 Subject: [tech] morwong's disks In-Reply-To: Message-ID: On Wed, 13 Aug 2003, Paul Day wrote: > On Wed, 13 Aug 2003, Paul Day wrote: > > > We've been damned lucky. Similar Seagate 9Gbs fail all the time at UCS > > > and we have an air conditioned machine room. I'd say go the 2 * RAID-1 > > > disk sets. Seems like the best option. > > > > Yeah, I've had similar problems with those 9GB SCSI Seagates. You can get > > them pretty cheap from the US though. > > Actually, James has just informed me that I actually provided one of them > - so for the reocrd, I claim no responsibility if it dies and strongly > recommend RAID1. ;) Good call: Initializing device dsk8. Initialization of disk device dsk8 failed. Error: lsm:voldisk: ERROR: Device dsk8: define failed: Disk read failure lsm:voldisk: ERROR: voldisk: Warning couldn't update disklabel on dsk8 Hit RETURN to continue. Goodbye. (dsk8 is said 9gig quantum viking II from Paul) Mustang: do you still have those SCA SBBs floating around? -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC President http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 / From grahame at angrygoats.net Thu Aug 14 21:29:10 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:51 2004 Subject: [tech] morwong's disks In-Reply-To: References: Message-ID: <1060867749.7638.40.camel@typhaon.ucs.uwa.edu.au> On Thu, 2003-08-14 at 21:26, James Andrewartha wrote: > On Wed, 13 Aug 2003, Paul Day wrote: > > > On Wed, 13 Aug 2003, Paul Day wrote: > > > > We've been damned lucky. Similar Seagate 9Gbs fail all the time at UCS > > > > and we have an air conditioned machine room. I'd say go the 2 * RAID-1 > > > > disk sets. Seems like the best option. > > > > > > Yeah, I've had similar problems with those 9GB SCSI Seagates. You can get > > > them pretty cheap from the US though. > > > > Actually, James has just informed me that I actually provided one of them > > - so for the reocrd, I claim no responsibility if it dies and strongly > > recommend RAID1. ;) > > Good call: > > Initializing device dsk8. > > Initialization of disk device dsk8 failed. > Error: lsm:voldisk: ERROR: Device dsk8: define failed: > Disk read failure > lsm:voldisk: ERROR: voldisk: Warning couldn't update disklabel on dsk8 > Hit RETURN to continue. > > Goodbye. > > (dsk8 is said 9gig quantum viking II from Paul) > > Mustang: do you still have those SCA SBBs floating around? James, Is it still in warranty? If it is, I can lend a 9Gb Seagate until a replacement turns up. Cheers Grahame From nick at ucc.gu.uwa.edu.au Thu Aug 14 22:09:17 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:51 2004 Subject: [tech] morwong's disks In-Reply-To: <1060867749.7638.40.camel@typhaon.ucs.uwa.edu.au> References: <1060867749.7638.40.camel@typhaon.ucs.uwa.edu.au> Message-ID: <20030814140916.GD2952@morwong.ucc.gu.uwa.edu.au> On Thu, Aug 14, 2003 at 09:29:10PM +0800, Grahame Bowland wrote: > Is it still in warranty? If it is, I can lend a 9Gb Seagate until a > replacement turns up. SCSI inquiry => QUANTUM VIKING II 9.1WLS199818941080 It's _just_ possible that it is - the manufacturer warranty was 5 years, so it it was bought after August 1998, there's a chance. Of course, if it was an OEM drive, or Maxtor don't want to speak to us (they bought Quantum's drive division) or it's going to take more time, effort, postage and packing than the market value of the drive (erm... $50?) then it won't be much help... Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From bernard at blackham.com.au Sun Aug 17 00:39:50 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:27:51 2004 Subject: [tech] the aftermath Message-ID: <20030816163950.GB5197@amidala> So around 9:44 am this morning, the power at UCC (and presumably most of campus went down). It came back about 9:49 am but with some vengeance... The flaky power froze the Cisco and the DECServer, powered down hydra, and frazzled meito's PSU and MB. All just needed some bouncing except meito... Skipping over the long and convoluted story, meito had it's drives transplanted to a new motherboard & CPU (mine) and seems to be running a little faster (IDE speed). The new CPU is a P3-450. The only downside is that there's no facility in the BIOS to tell it to switch on when it powers up - so it'll stay off in a power failure. But we're looking to fix that with a small circuit. At certain points during the day it was feared that both meito's disks had died horrible deaths, which proved quite distressing without a backup. (Even RAID1 wasn't enough to save them). Turns out it was the motherboards' at fault, but it still raises the question of backups at UCC... do we need a plan? I know some things get copied around at sporadic intervals, but should we have something more concrete? Bernard. -- Bernard Blackham bernard at blackham dot com dot au From nick at ucc.gu.uwa.edu.au Sun Aug 17 02:28:55 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:52 2004 Subject: [tech] the aftermath In-Reply-To: <20030816163950.GB5197@amidala> References: <20030816163950.GB5197@amidala> Message-ID: <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> On Sun, Aug 17, 2003 at 12:39:50AM +0800, Bernard Blackham wrote: [...] > Skipping over the long and convoluted story, meito had it's drives > transplanted to a new motherboard & CPU (mine) and seems to be > running a little faster (IDE speed). The new CPU is a P3-450. The So it is! 4.1MB/sec read speed, up from about 1MB/sec! > only downside is that there's no facility in the BIOS to tell it to > switch on when it powers up - so it'll stay off in a power failure. > But we're looking to fix that with a small circuit. We could just link the +5VSB line to the PS_ON line on the ATX connector. > At certain points during the day it was feared that both meito's > disks had died horrible deaths, which proved quite distressing > without a backup. (Even RAID1 wasn't enough to save them). Turns out > it was the motherboards' at fault, but it still raises the question > of backups at UCC... do we need a plan? I was thinking automated DVD image creation to mussel:/space . If a blank DVD is left in cybium overnight, it can ( fetch the next image && burn && delete it ). The slight problem is that Michal just just been burning some more DVDs and he's noticed that he gets occasional bit errors copying files around on cybium, which probably means the VIA DMA driving is FITH. It's not a complete killer because the backups would still be 99% OK, but still... cybium should probably get a new motherboard. ::-) In addition to anything else, we have a 20/40GB DLT and [GMB] says UCS have a spare 20/40GB DLT autoloader... Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From davyd at zdlcomputing.com Sun Aug 17 02:31:38 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:27:53 2004 Subject: [tech] the aftermath In-Reply-To: <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> References: <20030816163950.GB5197@amidala> <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> Message-ID: <1061058698.1245.1.camel@pingu> On Sun, 2003-08-17 at 02:28, Nick Bannon wrote: > In addition to anything else, we have a 20/40GB DLT and [GMB] says UCS > have a spare 20/40GB DLT autoloader... And it's entirely possibly I have some tapes lying around that are suitable. I picked up half a dozen free tapes recently... -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030817/accb0eb9/attachment.pgp From michal at tartarus.uwa.edu.au Sun Aug 17 02:42:49 2003 From: michal at tartarus.uwa.edu.au (Michal Gornisiewicz) Date: Wed Oct 27 01:27:53 2004 Subject: [tech] the aftermath In-Reply-To: <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> References: <20030816163950.GB5197@amidala> <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030816184248.GA15647@ucc.gu.uwa.edu.au> On Sun, Aug 17, 2003 at 02:28:55AM +0800, Nick Bannon wrote: > The slight problem is that Michal just just been burning some more DVDs > and he's noticed that he gets occasional bit errors copying files > around on cybium, which probably means the VIA DMA driving is FITH. It's > not a complete killer because the backups would still be 99% OK, but > still... cybium should probably get a new motherboard. ::-) It's not just _occasional_ bit errors. I couldn't copy a DVD image from my HDD to /space on cybium without error. I didn't investigate whether these were single bit errors like the dvd I burnt previously had, but md5sum checks for most of the files on the iso (on /space) were failing. I tried with & without DMA (first turned it off only for /space then both drives), so it's not a DMA issue, unless leaving DMA on for hda could've caused it which I doubt. Just reading off my HDD was fine (md5sums checked out). I ended up burning the dvd off my own HDD but I haven't looked at it yet. Will poke moro. MG -- "Reality is merely an illusion, albeit a very persistent one." -- Albert Einstein From bonfire at bur.st Sun Aug 17 08:57:06 2003 From: bonfire at bur.st (Paul Day) Date: Wed Oct 27 01:27:54 2004 Subject: [tech] the aftermath In-Reply-To: <20030816163950.GB5197@amidala> References: <20030816163950.GB5197@amidala> Message-ID: On Sun, 17 Aug 2003, Bernard Blackham wrote: > At certain points during the day it was feared that both meito's > disks had died horrible deaths, which proved quite distressing > without a backup. (Even RAID1 wasn't enough to save them). Turns out > it was the motherboards' at fault, but it still raises the question > of backups at UCC... do we need a plan? Old PC with a couple of cheap 120GB IDE drives, somewhere _other_ than UCC (ie, elsewhere on campus or colo-ed somewhere on WAIX for free) doing an rsync at 4am each morning? PD -- Paul Day Web: www.bur.st/~bonfire GPG Key ID: 2EF4ED23 From adrian at ucc.gu.uwa.edu.au Sun Aug 17 11:43:03 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:54 2004 Subject: [tech] the aftermath In-Reply-To: References: <20030816163950.GB5197@amidala> Message-ID: <20030817034303.GB9075@morwong.ucc.gu.uwa.edu.au> On Sun, Aug 17, 2003, Paul Day wrote: > Old PC with a couple of cheap 120GB IDE drives, somewhere _other_ than UCC > (ie, elsewhere on campus or colo-ed somewhere on WAIX for free) doing an > rsync at 4am each morning? .. once the network upgrade happens. It'll suck otherwise. What about a firewire caddy and a fat removeable hard disk? Adrian From nick at ucc.gu.uwa.edu.au Sun Aug 17 12:07:06 2003 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Wed Oct 27 01:27:54 2004 Subject: [tech] the aftermath In-Reply-To: <20030817034303.GB9075@morwong.ucc.gu.uwa.edu.au> References: <20030816163950.GB5197@amidala> <20030817034303.GB9075@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030817040706.GB29815@morwong.ucc.gu.uwa.edu.au> On Sun, Aug 17, 2003 at 11:43:03AM +0800, Adrian Chadd wrote: > On Sun, Aug 17, 2003, Paul Day wrote: > > Old PC with a couple of cheap 120GB IDE drives, somewhere _other_ than UCC > > (ie, elsewhere on campus or colo-ed somewhere on WAIX for free) doing an > > rsync at 4am each morning? > > .. once the network upgrade happens. It'll suck otherwise. > What about a firewire caddy and a fat removeable hard disk? Those are useful, but insufficient. They save us from "oh dear the hardware exploded", but they don't help for "someone fucked up the config/deleted something/rootkit'ed the machine but no-one noticed until a month later". This (and its links at the bottom) can help: http://www.mikerubel.org/computers/rsync_snapshots/ Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig@rcpt.to | I lack the time to make it shorter." - Pascal From bernard at blackham.com.au Sun Aug 17 12:28:02 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:27:55 2004 Subject: [tech] the aftermath In-Reply-To: <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> References: <20030816163950.GB5197@amidala> <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030817042802.GA8789@amidala> On Sun, Aug 17, 2003 at 02:28:55AM +0800, Nick Bannon wrote: > So it is! 4.1MB/sec read speed, up from about 1MB/sec! Using bonnie++ last night, I was recording read speeds of up to 20MB/s (sequential block reads), and write speeds of 9 MB/s. I've run some network speed benchmarks around the place and come up with the following: cybium -> velvet 11.00 MB/s cybium -> mussel 4.20 MB/s meito -> mussel 11.07 MB/s meito -> velvet 3.94 MB/s So anything that crosses hydra is bottlenecked at ~4MB/s anyway. How about that Accelar? :) Bernard. -- Bernard Blackham bernard at blackham dot com dot au From davyd at zdlcomputing.com Sun Aug 17 12:34:40 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:27:56 2004 Subject: [tech] the aftermath In-Reply-To: <20030817042802.GA8789@amidala> References: <20030816163950.GB5197@amidala> <20030816182855.GA8342@morwong.ucc.gu.uwa.edu.au> <20030817042802.GA8789@amidala> Message-ID: <1061094880.1246.7.camel@pingu> On Sun, 2003-08-17 at 12:28, Bernard Blackham wrote: > So anything that crosses hydra is bottlenecked at ~4MB/s anyway. How > about that Accelar? :) I think Grahame has it set up and ready to go. All that's required is to install *BSD on bananabox, get that sorted to do the upstream gateway (v4 and v6). Hydra can then be removed as is (in case of major fuckups), and the new hardware put in it's place. -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030817/39609bf3/attachment.pgp From rodney at ucc.gu.uwa.edu.au Sun Aug 17 16:45:26 2003 From: rodney at ucc.gu.uwa.edu.au (Rodney Lorrimar) Date: Wed Oct 27 01:27:56 2004 Subject: [tech] Debian mirror not updated? Message-ID: <20030817084526.GB1456@rodney.lorrimar.dropbear.id.au> Hello. I haven't had a Debian update for a few days now. It is making me NERVOUS I need my FIX! The file which ftp://ftp.uwa.edu.au/mirrors/linux/debian/dists/unstable/main/binary-i386/Packages refers to is dated Aug 7 but the corresponding file on PlanetMirror is dated yesterday Aug 16. Someone please HELP ME! From Rodney -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030817/509ba251/attachment.pgp From adrian at ucc.gu.uwa.edu.au Sun Aug 17 19:08:33 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:57 2004 Subject: [tech] Debian mirror not updated? In-Reply-To: <20030817084526.GB1456@rodney.lorrimar.dropbear.id.au> References: <20030817084526.GB1456@rodney.lorrimar.dropbear.id.au> Message-ID: <20030817110833.GC9075@morwong.ucc.gu.uwa.edu.au> Hm. Why not email support@ucs.uwa.edu.au ? UCC doesn't run the UWA FTP site. adrian (I'll bug grahame tomorrow about it, I don't know much about debian mirrors..) On Sun, Aug 17, 2003, Rodney Lorrimar wrote: > Hello. I haven't had a Debian update for a few days now. It is making > me NERVOUS I need my FIX! > > The file which > ftp://ftp.uwa.edu.au/mirrors/linux/debian/dists/unstable/main/binary-i386/Packages > refers to is dated Aug 7 but the corresponding file on PlanetMirror is > dated yesterday Aug 16. > > > Someone please HELP ME! > > > From Rodney From adrian at ucc.gu.uwa.edu.au Sun Aug 17 19:11:00 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:57 2004 Subject: [tech] the aftermath In-Reply-To: <20030817040706.GB29815@morwong.ucc.gu.uwa.edu.au> References: <20030816163950.GB5197@amidala> <20030817034303.GB9075@morwong.ucc.gu.uwa.edu.au> <20030817040706.GB29815@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030817111100.GD9075@morwong.ucc.gu.uwa.edu.au> On Sun, Aug 17, 2003, Nick Bannon wrote: > On Sun, Aug 17, 2003 at 11:43:03AM +0800, Adrian Chadd wrote: > > On Sun, Aug 17, 2003, Paul Day wrote: > > > Old PC with a couple of cheap 120GB IDE drives, somewhere _other_ than UCC > > > (ie, elsewhere on campus or colo-ed somewhere on WAIX for free) doing an > > > rsync at 4am each morning? > > > > .. once the network upgrade happens. It'll suck otherwise. > > What about a firewire caddy and a fat removeable hard disk? > > Those are useful, but insufficient. They save us from "oh dear the > hardware exploded", but they don't help for "someone fucked up the > config/deleted something/rootkit'ed the machine but no-one noticed > until a month later". Err, thats an application layer thing, _not_ the physical media. you can do incremental (dump, rsync, tar, etc) backup/restores quite happily onto hard disk. Adrian From grahame at angrygoats.net Mon Aug 18 23:47:43 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:57 2004 Subject: [tech] Hydra has been replaced by cthulhu Message-ID: <1061221663.1464.18.camel@solitaire> Hi all A quick update. Today hydra was replaced with a Accelar 1100B, namely cthulhu.ucc.gu.uwa.edu.au. The switch can be configured via telnet, SNMP and using the "dmgr" utility installed on cybium. This utility is quite unstable under windows 2000, so I'd suggest someone get a Win98 laptop or something to run it on. To run it; c: cd c:\dm c:\dm\bin\wish42 dmgr.tcl Password "uccpub" for read only, old root password for write community. There is a vlan trunk running into hydra; it is now running on 130.95.13.4 and 130.95.13.66 respectively. Whether it comes back or not after reboot has not been tested; it should do but please, be around to check if it doesn't... Many, many duplex issues were discovered, which have mostly been fixed. Here is Grahame's Helpful Guide to Duplex Negotiation: 1. if remote host is an unmanaged switch or a server, always true autonegotiation first. if you turn autonegotiation off, many switches will default to 100/half. 2. if that doesn't work, try forcing to 100/full. 3. if that doesn't work, curse and set your end to 100/half too. Most of the little switches were at 100/half when we were forced 100/full (left over config from when the thing was at UCS) so that needed fixing. This caused the initial (dismal) results of perfomance testing - we are now seeing >10MB/sec between subnets (vlans). *under no circumstances turn the webserver on* It is buggy, and code red causes the switch to slow down / lock up. This is a known problem, possibly fixed with later code - we should probably get some from somewhere. But there is no reason to turn it on. Play with the switch, but please, please don't do something awful to it so we can't telnet in. The serial is a PITA :) Guild network upgrade status; fibre is in, change-over to 100Mb will hopefully happen tomorrow or the day after. At which point, connectivity to UWA should get massively faster. Thanks in particular should go to James, for running around madly fixing duplex problems :-) -- Grahame Bowland doth safely trust in her so that he could not be eaten they are so From grahame at angrygoats.net Tue Aug 19 00:45:18 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:57 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <20030818161510.GA13371@amidala> References: <20030818161510.GA13371@amidala> Message-ID: <1061225118.1583.26.camel@solitaire> On Tue, 2003-08-19 at 00:15, Bernard Blackham wrote: > Also UCC's core router just changed, so ssh.ucc.asn.au isn't > currently working either. If you can use scp, that'll be working > again soon (tomorrow ideally :) AFAIK the best way to set up ssh.ucc.asn.au and so on is to add interfaces to hydra, AKA eth3.2:0, eth3.2:1, with the IPs of ssh.ucc.asn.au and flame.ucc.asn.au, and then do the NAT magic onto them :-) Should work well enough. In other news, I've got deborah@ucs chasing up LSM license for Tru64 - the thing told us to fark off when trs80 tried to make a mirrored set, and it turns out LSM was last seen in the 2000 license PAK. It might be that UCS ended up with different terms from DEC/Compaq/HP, but I doubt it. I looked at DECUS hobbyist licenses, but they suck for Tru64 - AU$200 for a license that wouldn't include LSM anyway. OpenVMS is free for all features as far as I can see, so if anyone wants to install it somewhere I have media and a license file. -- Grahame Bowland sons to be bondmen and elisha said unto him get thee from me From bernard at blackham.com.au Tue Aug 19 00:56:04 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:27:57 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <1061225118.1583.26.camel@solitaire> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> Message-ID: <20030818165604.GA14136@amidala> On Tue, Aug 19, 2003 at 12:45:18AM +0800, Grahame Bowland wrote: > On Tue, 2003-08-19 at 00:15, Bernard Blackham wrote: > > Also UCC's core router just changed, so ssh.ucc.asn.au isn't > > currently working either. If you can use scp, that'll be working > > again soon (tomorrow ideally :) > > AFAIK the best way to set up ssh.ucc.asn.au and so on is to add > interfaces to hydra, AKA eth3.2:0, eth3.2:1, with the IPs of > ssh.ucc.asn.au and flame.ucc.asn.au, and then do the NAT magic onto them > :-) Sounds good - will try it in the morning. So far as I can tell, the Accelar can't do ip aliasing - so the 172.26. gateways (.3, .65, etc) will probably have to be a hydra thing, when WAIX comes back. Bernard. -- Bernard Blackham bernard at blackham dot com dot au From grahame at angrygoats.net Tue Aug 19 02:44:10 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:58 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <20030818165604.GA14136@amidala> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> Message-ID: <1061232250.1582.38.camel@solitaire> On Tue, 2003-08-19 at 00:56, Bernard Blackham wrote: > On Tue, Aug 19, 2003 at 12:45:18AM +0800, Grahame Bowland wrote: > > On Tue, 2003-08-19 at 00:15, Bernard Blackham wrote: > > > Also UCC's core router just changed, so ssh.ucc.asn.au isn't > > > currently working either. If you can use scp, that'll be working > > > again soon (tomorrow ideally :) > > > > AFAIK the best way to set up ssh.ucc.asn.au and so on is to add > > interfaces to hydra, AKA eth3.2:0, eth3.2:1, with the IPs of > > ssh.ucc.asn.au and flame.ucc.asn.au, and then do the NAT magic onto them > > :-) > > Sounds good - will try it in the morning. > > So far as I can tell, the Accelar can't do ip aliasing - so the > 172.26. gateways (.3, .65, etc) will probably have to be a hydra > thing, when WAIX comes back. Ah, that's going to suck. I'm sure there must be a way to do it. Otherwise, people leaching off WAIX are going to get hit with hydra speeds. I suspect one way to do it is to create another VLAN with the 172.26.42.3 IP on it, and then add it to all of the machine room ports, without selecting trunking. Try that, it might work (obviously, try it on one port, like mojarra, see if it works, then go on from there :P) It makes sense for that to work, anyway. I'd poke it, but it's not a good idea from here :-) -- Grahame Bowland finger some of the people that walked in darkness have seen a great light From adrian at ucc.gu.uwa.edu.au Tue Aug 19 10:29:11 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:58 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <1061232250.1582.38.camel@solitaire> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> <1061232250.1582.38.camel@solitaire> Message-ID: <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> On Tue, Aug 19, 2003, Grahame Bowland wrote: > > So far as I can tell, the Accelar can't do ip aliasing - so the > > 172.26. gateways (.3, .65, etc) will probably have to be a hydra > > thing, when WAIX comes back. > > Ah, that's going to suck. I'm sure there must be a way to do it. > Otherwise, people leaching off WAIX are going to get hit with hydra > speeds. Hey, that might be a good thing, thankyou very much. Adrian From dayta at ucc.gu.uwa.edu.au Tue Aug 19 10:47:00 2003 From: dayta at ucc.gu.uwa.edu.au (Leighton Haynes) Date: Wed Oct 27 01:27:58 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> <1061232250.1582.38.camel@solitaire> <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030819024700.GI19980@ucc.gu.uwa.edu.au> On Tue, Aug 19, 2003 at 10:29:11AM +0800, Adrian Chadd wrote: > On Tue, Aug 19, 2003, Grahame Bowland wrote: > > > > So far as I can tell, the Accelar can't do ip aliasing - so the > > > 172.26. gateways (.3, .65, etc) will probably have to be a hydra > > > thing, when WAIX comes back. > > > > Ah, that's going to suck. I'm sure there must be a way to do it. > > Otherwise, people leaching off WAIX are going to get hit with hydra > > speeds. > > Hey, that might be a good thing, thankyou very much. I was thinking the very same thing. Leighton... -- #0421 113 305 - dayta@ucc.gu.uwa.edu.au How do you expect?/I will know what to do./When all I know. Is what you tell me to. - Linkin Park/By Myself From davyd at zdlcomputing.com Tue Aug 19 10:50:47 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:27:59 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <20030819024700.GI19980@ucc.gu.uwa.edu.au> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> <1061232250.1582.38.camel@solitaire> <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> <20030819024700.GI19980@ucc.gu.uwa.edu.au> Message-ID: <1061261447.2088.1.camel@pingu> On Tue, 2003-08-19 at 10:47, Leighton Haynes wrote: > I was thinking the very same thing. Or we get rid of the WAIX IPs. -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030819/e6dfe947/attachment.pgp From dayta at ucc.gu.uwa.edu.au Tue Aug 19 10:55:57 2003 From: dayta at ucc.gu.uwa.edu.au (Leighton Haynes) Date: Wed Oct 27 01:27:59 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <1061261447.2088.1.camel@pingu> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> <1061232250.1582.38.camel@solitaire> <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> <20030819024700.GI19980@ucc.gu.uwa.edu.au> <1061261447.2088.1.camel@pingu> Message-ID: <20030819025557.GJ19980@ucc.gu.uwa.edu.au> On Tue, Aug 19, 2003 at 10:50:47AM +0800, Davyd wrote: > On Tue, 2003-08-19 at 10:47, Leighton Haynes wrote: > > > I was thinking the very same thing. > > Or we get rid of the WAIX IPs. That has valid uses for, say, me reading my email ;) And playing games. :P Leighton... -- #0421 113 305 - dayta@ucc.gu.uwa.edu.au How do you expect?/I will know what to do./When all I know. Is what you tell me to. - Linkin Park/By Myself From adrian at ucc.gu.uwa.edu.au Tue Aug 19 11:04:13 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:27:59 2004 Subject: [tech] Re: [wheel] Ftping to UCC servers In-Reply-To: <1061261447.2088.1.camel@pingu> References: <20030818161510.GA13371@amidala> <1061225118.1583.26.camel@solitaire> <20030818165604.GA14136@amidala> <1061232250.1582.38.camel@solitaire> <20030819022911.GA85035@morwong.ucc.gu.uwa.edu.au> <20030819024700.GI19980@ucc.gu.uwa.edu.au> <1061261447.2088.1.camel@pingu> Message-ID: <20030819030413.GB85035@morwong.ucc.gu.uwa.edu.au> On Tue, Aug 19, 2003, Davyd wrote: > On Tue, 2003-08-19 at 10:47, Leighton Haynes wrote: > > > I was thinking the very same thing. > > Or we get rid of the WAIX IPs. Got a better idea? Adrian From grahame at angrygoats.net Tue Aug 19 13:31:43 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:59 2004 Subject: [tech] What is the setup on hydra Message-ID: <1061271103.29431.44.camel@typhaon.ucs.uwa.edu.au> And did anyone try my idea of setting up a new VLAN on the Accelar and doing the routing for the 172 addresses there? It would be really good if people posted the changes to the configuration to the list. For example: hydra seems to have stopped running a .1q trunk and has been plugged into all the VLANs (good, this seems to have fixed the .4 / .66 problem) hydra now seems to be routing the 172 addresses (bad, this should be possible to do a lot better) Anyone? Cheers Grahame From grahame at angrygoats.net Tue Aug 19 15:31:40 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:27:59 2004 Subject: [tech] Updates Message-ID: <1061278300.29431.51.camel@typhaon.ucs.uwa.edu.au> Ok, I discovered some of the problems with the accelar management software is that we're using 1.0. 1.3.5 is on a laptop at work, I'm going to copy it over and get it going under wine. The newer version supports setting the security stuff right. I've set up the rw, rwa logins to be the old root password, ro is "uccpub". From grahame at angrygoats.net Tue Aug 19 15:42:25 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:28:00 2004 Subject: [tech] Newer version of DM Message-ID: <1061278944.29240.54.camel@typhaon.ucs.uwa.edu.au> Do not use the old c:\dm install on cybium. It *will* break things. The zip at http://typhaon.ucs.uwa.edu.au/~grahame/Dm.zip needs to be unziped to c:\dm (move the old one first) Run it with cd c: cd c:\dm c:\dm\bin\wish42.exe dmgr.tcl This utility is pretty damned useful, and the newer version seems stable and allows configuring a whole bunch of stuff. Toivo has told me how to get the multiple IP address stuff going, so I'll do that tonight. From bernard at blackham.com.au Wed Aug 20 11:52:17 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:28:00 2004 Subject: [tech] Newer version of DM In-Reply-To: <1061278944.29240.54.camel@typhaon.ucs.uwa.edu.au> References: <1061278944.29240.54.camel@typhaon.ucs.uwa.edu.au> Message-ID: <20030820035217.GA1035@amidala> On Tue, Aug 19, 2003 at 03:42:25PM +0800, Grahame Bowland wrote: > The zip at http://typhaon.ucs.uwa.edu.au/~grahame/Dm.zip needs to be > unziped to c:\dm (move the old one first) I've installed it on Cybium. > This utility is pretty damned useful, and the newer version seems stable > and allows configuring a whole bunch of stuff. Agreed - it works that much better :) > Toivo has told me how to get the multiple IP address stuff going, so > I'll do that tonight. I haven't got multiple IPs going, but I've created extra VLANs "byIpSubnet" for machineroom_waix and clubroom_waix, restricted to the relative ports. All seems sane and peachy. I'll do some more testing this arvo. Bernard. -- Bernard Blackham bernard at blackham dot com dot au From alastair at ucc.gu.uwa.edu.au Thu Aug 21 23:36:44 2003 From: alastair at ucc.gu.uwa.edu.au (Alastair Irvine) Date: Wed Oct 27 01:28:00 2004 Subject: [tech] WAIX connectivity Message-ID: <20030821153644.GA24050@ucc.gu.uwa.edu.au> Hi all. Anyone know when UCC will be reachable via its WAIX IPs? Thanks. Ciao -- ... Most of our lives are about proving something, either to ourselves or to someone else. _____________________________________________________________________ | | | -=*Alastair Irvine*=- | | C-monkey/wanderer/board&RPGer/net-nut alastair@ucc.gu.uwa.edu.au | |_____________________________________________________________________| From davyd at zdlcomputing.com Fri Aug 22 07:35:28 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:28:01 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030821153644.GA24050@ucc.gu.uwa.edu.au> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> Message-ID: <1061508928.1335.0.camel@pingu> On Thu, 2003-08-21 at 23:36, Alastair Irvine wrote: > Hi all. Anyone know when UCC will be reachable via its WAIX IPs? At current speed... when Aarnet3 is deployed and the firewalls can be dropped. -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030822/855cd21b/attachment.pgp From acolyte at ucc.gu.uwa.edu.au Fri Aug 22 07:46:08 2003 From: acolyte at ucc.gu.uwa.edu.au (Andrew Bailey) Date: Wed Oct 27 01:28:01 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030821153644.GA24050@ucc.gu.uwa.edu.au> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> Message-ID: <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> On Thu, Aug 21, 2003 at 11:36:44PM +0800, Alastair Irvine wrote: > Hi all. Anyone know when UCC will be reachable via its WAIX IPs? > Hi all. Anyone know when UCC will be reacable via ssh reliably. Andrew -- "The hot dog eating contest is not only a beautiful display of athleticism, it is a fundamental way for citizens of all nations to display patriotism," - Wayne Norbitz From davyd at zdlcomputing.com Fri Aug 22 07:49:38 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> Message-ID: <1061509778.1334.2.camel@pingu> On Fri, 2003-08-22 at 07:46, Andrew Bailey wrote: > Hi all. Anyone know when UCC will be reacable via ssh reliably. Does ssh.ucc.asn.au work? Failing that, see previous note... failing that, wait until infinity centre's on the UCC machine room. -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030822/6bbf49a9/attachment.pgp From acolyte at ucc.gu.uwa.edu.au Fri Aug 22 07:58:40 2003 From: acolyte at ucc.gu.uwa.edu.au (Andrew Bailey) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: <1061509778.1334.2.camel@pingu> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> Message-ID: <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> On Fri, Aug 22, 2003 at 07:49:38AM +0800, Davyd wrote: > On Fri, 2003-08-22 at 07:46, Andrew Bailey wrote: > > > Hi all. Anyone know when UCC will be reacable via ssh reliably. > > Does ssh.ucc.asn.au work? no. Though ssh.ucc.gu.uwa.edu.au might, I haven't tried that. > > Failing that, see previous note... failing that, wait until infinity > centre's on the UCC machine room. > Nah, I'll just wait until it pisses off a competent wheel member :) Andrew. -- "The hot dog eating contest is not only a beautiful display of athleticism, it is a fundamental way for citizens of all nations to display patriotism," - Wayne Norbitz From bernard at blackham.com.au Fri Aug 22 09:15:01 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> Message-ID: <20030822011501.GA10883@amidala> On Fri, Aug 22, 2003 at 07:58:40AM +0800, Andrew Bailey wrote: > > > Hi all. Anyone know when UCC will be reacable via ssh reliably. > > > > Does ssh.ucc.asn.au work? > > no. Though ssh.ucc.gu.uwa.edu.au might, I haven't tried that. Should be the same thing. And they should both be working. They were down for a day or three when the core router was changed one night, and a guild tech accidentally dropped UCC's uplink the next night. Bernard. -- Bernard Blackham bernard at blackham dot com dot au From dave at difference.com.au Fri Aug 22 09:22:17 2003 From: dave at difference.com.au (David Cake) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030822011501.GA10883@amidala> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> <20030822011501.GA10883@amidala> Message-ID: > They were >down for a day or three when the core router was changed one night, >and a guild tech accidentally dropped UCC's uplink the next night. Huh what? We did have a switch that wedged itself one night, yes. But I don't think the Guild has ever had anyone involved in the networking side who hasn't been an UCC member at some point, and certainly we've tried pretty hard to preserve UCCs connectivity and its fared no differently to anyone else. And spent a lot of time consulting with UCC over connectivity options, choice of equipment, etc. I rather resent the implication that mysterious guild techs screw the UCC around. Cheers David From dayta at ucc.gu.uwa.edu.au Fri Aug 22 09:25:23 2003 From: dayta at ucc.gu.uwa.edu.au (Leighton Haynes) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> <20030822011501.GA10883@amidala> Message-ID: <20030822012522.GC21170@ucc.gu.uwa.edu.au> On Fri, Aug 22, 2003 at 09:22:17AM +0800, David Cake wrote: > > They were > >down for a day or three when the core router was changed one night, > >and a guild tech accidentally dropped UCC's uplink the next night. > > Huh what? > We did have a switch that wedged itself one night, yes. > But I don't think the Guild has ever had anyone involved in > the networking side who hasn't been an UCC member at some point, and > certainly we've tried pretty hard to preserve UCCs connectivity and > its fared no differently to anyone else. And spent a lot of time > consulting with UCC over connectivity options, choice of equipment, > etc. Yes, and we appreciate it. > I rather resent the implication that mysterious guild techs > screw the UCC around. I was wondering who he was referring to myself ;) I just assumed it was some random Guild hack fiddling. I'm sure no slander was intended. Leighton... -- #0421 113 305 - dayta@ucc.gu.uwa.edu.au How do you expect?/I will know what to do./When all I know. Is what you tell me to. - Linkin Park/By Myself From bernard at blackham.com.au Fri Aug 22 09:28:11 2003 From: bernard at blackham.com.au (Bernard Blackham) Date: Wed Oct 27 01:28:02 2004 Subject: [tech] WAIX connectivity In-Reply-To: References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> <20030822011501.GA10883@amidala> Message-ID: <20030822012811.GB10883@amidala> On Fri, Aug 22, 2003 at 09:22:17AM +0800, David Cake wrote: > > They were > >down for a day or three when the core router was changed one night, > >and a guild tech accidentally dropped UCC's uplink the next night. > > I rather resent the implication that mysterious guild techs > screw the UCC around. Sorry, I wasn't implying that at all. I know it wasn't at all deliberate. And it was fixed as soon as CH opened the next morning (thank you! :) Apologies for any offence. Bernard. -- Bernard Blackham bernard at blackham dot com dot au From davyd at zdlcomputing.com Fri Aug 22 09:36:53 2003 From: davyd at zdlcomputing.com (Davyd) Date: Wed Oct 27 01:28:03 2004 Subject: [tech] WAIX connectivity In-Reply-To: References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> <20030821234608.GA205804@morwong.ucc.gu.uwa.edu.au> <1061509778.1334.2.camel@pingu> <20030821235840.GB205804@morwong.ucc.gu.uwa.edu.au> Message-ID: <1061516212.1292.7.camel@pingu> Apologising on behalf of Bernard (who actually just posted): On Fri, 2003-08-22 at 09:22, David Cake wrote: > I rather resent the implication that mysterious guild techs > screw the UCC around. David, He wasn't implying it was your fault. Noone has had anything but praise for the virtual flawlessness with which the network upgrade has been carried out. My personal congratulations on avoiding any major disasters knocking us off the air for a week. The loss in connectivity was actually blamed on UCC hardware, rather then it being a fault of the guild. When we then discovered it was that a switch had simply crashed, we were actually impressed with the responsiveness that it was got back on the air. Personally, I think the Guild techs do really well by us. I don't think it's ever been suggested that they screw us around. There's only the occasional joke about someone tripping over "The Big Blue Cable". Once again, I hope you understand Bernard intended no offense, and I feel he too would in fact congratulate you, as I have, on the fantastic job you and the other technicians involved did on upgrading the network. Regards --Davyd Madeley Committee, 2003 -- http://davyd.ucc.asn.au/ PGP Fingerprint 08B0 341A 0B9B 08BB 2118 C060 2EDD BB4F 5191 6CDA -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 227 bytes Desc: This is a digitally signed message part Url : http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20030822/67039a8d/attachment.pgp From adrian at ucc.gu.uwa.edu.au Fri Aug 22 10:10:29 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:28:03 2004 Subject: [tech] WAIX connectivity In-Reply-To: <20030821153644.GA24050@ucc.gu.uwa.edu.au> References: <20030821153644.GA24050@ucc.gu.uwa.edu.au> Message-ID: <20030822021029.GA172508@morwong.ucc.gu.uwa.edu.au> On Thu, Aug 21, 2003, Alastair Irvine wrote: > Hi all. Anyone know when UCC will be reachable via its WAIX IPs? When hell freezes over[1]. I've been pondering many different methods of gaining access to UCC via WAIX but they all revolve around a tunnel to insure the traffic is _always free_. Adrian [1] Yes, when aarnet3 comes out and somehow the pricing structure for departments changes. There's no guarantee on the latter. From alastair at ucc.gu.uwa.edu.au Fri Aug 22 11:59:09 2003 From: alastair at ucc.gu.uwa.edu.au (Alastair Irvine) Date: Wed Oct 27 01:28:04 2004 Subject: [tech] From: header in mutt Message-ID: <20030822035909.GA23959@ucc.gu.uwa.edu.au> Hi all. Any idea why {mussel,mermaid}:/etc/Muttrc has the following options? They seem kind of pointless to me. # don't add the hostname to the From header unset use_domain # don't generate a From header unset use_from -- ... "Censorship can't eliminate evil, it can only kill freedom" _____________________________________________________________________ | | | -=*Alastair Irvine*=- | | C-monkey/wanderer/board&RPGer/net-nut alastair@ucc.gu.uwa.edu.au | |_____________________________________________________________________| From adrian at ucc.gu.uwa.edu.au Fri Aug 22 12:04:26 2003 From: adrian at ucc.gu.uwa.edu.au (Adrian Chadd) Date: Wed Oct 27 01:28:04 2004 Subject: [tech] From: header in mutt In-Reply-To: <20030822035909.GA23959@ucc.gu.uwa.edu.au> References: <20030822035909.GA23959@ucc.gu.uwa.edu.au> Message-ID: <20030822040426.GE172508@morwong.ucc.gu.uwa.edu.au> On Fri, Aug 22, 2003, Alastair Irvine wrote: > Hi all. Any idea why {mussel,mermaid}:/etc/Muttrc has the following > options? They seem kind of pointless to me. > > # don't add the hostname to the From header > unset use_domain > # don't generate a From header > unset use_from Because your MTA can add them for you. I override them anyway. Adrian From trent at ucc.gu.uwa.edu.au Sun Aug 31 00:09:45 2003 From: trent at ucc.gu.uwa.edu.au (Trent Lloyd) Date: Wed Oct 27 01:28:04 2004 Subject: [tech] doh Message-ID: <20030830160945.GA9186@ucc.gu.uwa.edu.au> Heyas, if anyone with machine room keys gets into UCC tomorow can you poke seven - i think its marked "ipv6.ztsoftware.net" if you dont know which box it is? it appears to have crashed.. while not doing much. a reset would be good Cheers, Trent Sixlabs From grahame at angrygoats.net Sun Aug 31 12:54:36 2003 From: grahame at angrygoats.net (Grahame Bowland) Date: Wed Oct 27 01:28:04 2004 Subject: [tech] doh In-Reply-To: <20030830160945.GA9186@ucc.gu.uwa.edu.au> References: <20030830160945.GA9186@ucc.gu.uwa.edu.au> Message-ID: <20030831045436.GA9203@angrygoats.net> On Sun, Aug 31, 2003 at 12:09:45AM +0800, Trent Lloyd wrote: > Heyas, > > if anyone with machine room keys gets into UCC tomorow can you poke > seven - i think its marked "ipv6.ztsoftware.net" if you dont know which > box it is? > > it appears to have crashed.. while not doing much. > > a reset would be good You know what- you of all people, really really need a watchdog timer card. From trent at ucc.gu.uwa.edu.au Sun Aug 31 17:45:40 2003 From: trent at ucc.gu.uwa.edu.au (Trent Lloyd) Date: Wed Oct 27 01:28:04 2004 Subject: [tech] Re: [hwc] Upgrades (from other lists) In-Reply-To: <20030831045747.GA9352@angrygoats.net> References: <1062263671.1424.20.camel@pingu> <20030831045747.GA9352@angrygoats.net> Message-ID: <20030831094540.GB513245@morwong.ucc.gu.uwa.edu.au> On Sun, Aug 31, 2003 at 12:57:47PM +0800, Grahame Bowland wrote: > On Sun, Aug 31, 2003 at 01:14:35AM +0800, Davyd wrote: > > There has been talk on the lists about upgrading mermaid. > > > > Discussion with others has come up with some suggestions: > > - replace mermaid with an AMD (similar to pitch and velvet) > > - replace mermaid with something like an X-box > > - replace mussel with something REAL CHUNKY > > (such as a dual p4-xeon or dual-g5), leaving old-mussel to do > > mermaids tasks. > > Ugh, dual-p4 is boring :) dual-g5 possibly but rather pricey. Dual > opteron is worth looking into, I don't think it'd be *that* expensive. > > > It also occurs to me that mojarra no longer appears to have any tasks. > > In my opinion it could probably be amalgamated with mermaid, when > > mermaid is replaced. > > mojarra is a WAIX only box. Once I get the WAIX stuff sorted (hopefully > this week) it'll be useful. dont we have BGP routes on mussel ? > > > Personally I'm all for the last option, mussel would get faster, as well > > as allowing us to acquire another fast user shell machine. This means > > that morwong logins could be turned off (making morwong not suck) and we > > could get rid of mojarra. > > Don't turn logins off on morwong, it's got really really nice > programming manpages. :) > > > Also, a dual-g5 (which I think is available in a rackable version) would > > sort the men out from the boys, people would have to cross-compile for > > x86 ;) > > From trs80 at ucc.gu.uwa.edu.au Sun Aug 31 18:14:47 2003 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Wed Oct 27 01:28:05 2004 Subject: [tech] Re: [hwc] Upgrades (from other lists) In-Reply-To: <20030831094540.GB513245@morwong.ucc.gu.uwa.edu.au> Message-ID: On Sun, 31 Aug 2003, Trent Lloyd wrote: > On Sun, Aug 31, 2003 at 12:57:47PM +0800, Grahame Bowland wrote: > > On Sun, Aug 31, 2003 at 01:14:35AM +0800, Davyd wrote: > > > It also occurs to me that mojarra no longer appears to have any tasks. > > > In my opinion it could probably be amalgamated with mermaid, when > > > mermaid is replaced. > > > > mojarra is a WAIX only box. Once I get the WAIX stuff sorted (hopefully > > this week) it'll be useful. > > dont we have BGP routes on mussel ? Not at the moment - with 172 IPs not working, there's no point having BGP on. More to the point, mojarra will always work for WAIX stuff, regardless of whether zebra gets upgraded to a version with subtly different syntax. -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC President http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 /