From harrymc at ucc.gu.uwa.edu.au Tue Feb 1 14:47:36 2011 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Tue, 01 Feb 2011 14:47:36 +0800 Subject: [tech] Fwd: Radiotelescope project Message-ID: <4D47AC88.4050606@ucc.gu.uwa.edu.au> Bob suggested forwarding to tech. -------- Original Message -------- Subject: Radiotelescope project Date: Tue, 01 Feb 2011 14:43:43 +0800 From: Harry To: committee at ucc.gu.uwa.edu.au Hello Committee I'm about to build up an interface board (shield) between an Arduino Mega and the two stepper driver modules for the Radiotelescope. I've got some stepper motor code done on the Arduino which could be used to serially instruct the stepping (ramp up, constant speed slew, ramp down) in the final system. Alternatively, it could be used for testing the mount motion and then reassigned to another project requiring an Arduino. The parts cost (including Arduino Mega) is around $150 and I've just asked Bob whether that is appropriate expenditure to be reimbursed as a UCC project. If UCC has an Arduino Mega then the interface-board-only cost is closer to $70. I can buy the parts and proceed but if it runs and remains at UCC, would committee let me know if that is a recoverable cost ? All the best Harry From zanchey at ucc.gu.uwa.edu.au Tue Feb 1 21:42:35 2011 From: zanchey at ucc.gu.uwa.edu.au (David Adam) Date: Tue, 1 Feb 2011 21:42:35 +0800 (WST) Subject: [tech] Process ulimits on Mussel and Martello Message-ID: So a certain genius who will remain nameless decided to see if a forkbomb would work on Mussel. Apparently this is no longer deserving of an account locking, but as it's not the first time in recent years I decided it was probably time we did something about it. /etc/security/limits.conf on Mussel and Martello has been set with a soft limit of 4096 processes on all user accounts. getrlimit(2) informs me that on Linux this enforces a limit of 4096 threads per real UID. That's still enough to build Mozilla Firefox and run my screen session, and it's a soft limit anyway so if you're really struggling you can just bump it up with `ulimit -u onezillion` or whatever. 4096 was a number I pulled out of the air; there is little to no science behind it and is not intended to stand up to malicious attacks. There are still at least a thousand ways of exhausting resources on multiuser Linux systems anyway. David Adam UCC Wheel Member zanchey at ucc.gu.uwa.edu.au From dunc+wheel at dunc.org Tue Feb 1 22:34:22 2011 From: dunc+wheel at dunc.org (Duncan Sargeant) Date: Tue, 1 Feb 2011 15:34:22 +0100 Subject: [tech] Process ulimits on Mussel and Martello In-Reply-To: References: Message-ID: Proper engineer practice calls for you to use a fork bomb to refine the limit to a suitable value ... *grin* ,dunc On 1 February 2011 14:42, David Adam wrote: > So a certain genius who will remain nameless decided to see if a forkbomb > would work on Mussel. Apparently this is no longer deserving of an account > locking, but as it's not the first time in recent years I decided it was > probably time we did something about it. > > /etc/security/limits.conf on Mussel and Martello has been set with a soft > limit of 4096 processes on all user accounts. getrlimit(2) informs me that > on Linux this enforces a limit of 4096 threads per real UID. That's still > enough to build Mozilla Firefox and run my screen session, and it's a soft > limit anyway so if you're really struggling you can just bump it up with > `ulimit -u onezillion` or whatever. > > 4096 was a number I pulled out of the air; there is little to no science > behind it and is not intended to stand up to malicious attacks. There are > still at least a thousand ways of exhausting resources on multiuser Linux > systems anyway. > > David Adam > UCC Wheel Member > zanchey at ucc.gu.uwa.edu.au > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20110201/8a59acf0/attachment.htm From Adrian at Diskworld.com.au Thu Feb 3 13:42:50 2011 From: Adrian at Diskworld.com.au (Adrian Woodley) Date: Thu, 03 Feb 2011 13:42:50 +0800 Subject: [tech] Process ulimits on Mussel and Martello In-Reply-To: References: Message-ID: <4D4A405A.5080701@Diskworld.com.au> Heaven help us if medical doctors start using engineering practice. On 02/01/11 22:34, Duncan Sargeant wrote: > Proper engineer practice calls for you to use a fork bomb to refine > the limit to a suitable value ... *grin* > > ,dunc > > On 1 February 2011 14:42, David Adam > wrote: > > So a certain genius who will remain nameless decided to see if a > forkbomb > would work on Mussel. Apparently this is no longer deserving of an > account > locking, but as it's not the first time in recent years I decided > it was > probably time we did something about it. > > /etc/security/limits.conf on Mussel and Martello has been set with > a soft > limit of 4096 processes on all user accounts. getrlimit(2) informs > me that > on Linux this enforces a limit of 4096 threads per real UID. > That's still > enough to build Mozilla Firefox and run my screen session, and > it's a soft > limit anyway so if you're really struggling you can just bump it > up with > `ulimit -u onezillion` or whatever. > > 4096 was a number I pulled out of the air; there is little to no > science > behind it and is not intended to stand up to malicious attacks. > There are > still at least a thousand ways of exhausting resources on > multiuser Linux > systems anyway. > > David Adam > UCC Wheel Member > zanchey at ucc.gu.uwa.edu.au > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.ucc.gu.uwa.edu.au/pipermail/tech/attachments/20110203/9de423e2/attachment.htm From bob at ucc.gu.uwa.edu.au Fri Feb 4 23:09:24 2011 From: bob at ucc.gu.uwa.edu.au (Bob Adamson) Date: Fri, 4 Feb 2011 23:09:24 +0800 (WST) Subject: [tech] Motsugo root fs disk speeds Message-ID: Hi, I did a 'draft' debian install on motsugo (the new martello) today, and thought I might run bonnie++ to test the ssd disk speeds before I wiped it. Here are the results for the curious: Version 1.96 ------Sequential Output------ --Sequential Input- --Random- Concurrency 1 -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks-- Machine Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP /sec %CP motsugo 24G 717 99 255174 40 105141 13 4559 99 302405 16 3178 93 Latency 17676us 601ms 400ms 2069us 197ms 98370us Version 1.96 ------Sequential Create------ --------Random Create-------- motsugo -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete-- files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP 16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ Latency 324us 533us 552us 391us 25us 51us Looks like money well spent! No ETA on when everything will be properly online yet... Bob Adamson UCC Treasurer |"Bureaucracy is a challenge to the be conquered with a righteous | |attitude, an intolerance for stupidity, and a bulldozer when necessary" | | ---Peter's Laws | From frenchie at ucc.gu.uwa.edu.au Sat Feb 12 17:21:40 2011 From: frenchie at ucc.gu.uwa.edu.au (James French) Date: Sat, 12 Feb 2011 17:21:40 +0800 Subject: [tech] Pinball upgraded to 10.10 Message-ID: Hi All, I got sick of pinball not accepting auth reliably, bumped it up, seems to be better. Please report any problems.\ Regards, F. From maset at ucc.asn.au Sun Feb 13 12:47:27 2011 From: maset at ucc.asn.au (Anil Sharma) Date: Sun, 13 Feb 2011 12:47:27 +0800 Subject: [tech] [wheel] Committee Meeting This Saturday In-Reply-To: References: Message-ID: There seems to be some confusion between the roles of wheel and tech. This is a tech issue. At worst it requires committee. From matt at ucc.asn.au Mon Feb 21 19:33:17 2011 From: matt at ucc.asn.au (Matt Johnston) Date: Mon, 21 Feb 2011 19:33:17 +0800 Subject: [tech] Dovecot IMAP/POP3 on motsugo Message-ID: <20110221113317.GG24077@ucc.gu.uwa.edu.au> Hi, I've installed dovecot2 on motsugo and pointed secure.ucc and mussel's imapproxy at it, since that's where local mail lives now. Dovecot2 has the benefit of supporting IMAP COMPRESS which a few smartphone clients can use. It's installed in /usr/local/dovecot2 with a startup file in /etc/init.d. Nothing seems broken that I can see. Matt From bob at ucc.gu.uwa.edu.au Tue Feb 22 12:16:58 2011 From: bob at ucc.gu.uwa.edu.au (Bob Adamson) Date: Tue, 22 Feb 2011 12:16:58 +0800 (WST) Subject: [tech] Serial Thingemy-Whatsit Message-ID: The serial->ethernet server which connects the serial terminals in the corridor has broken. It's a Lantronix ets8uf, and it's blowing fuses immediately when power is supplied. I've pulled the cover off, but nothing jumps out at me as broken. Anyway...does anyone know what we might be able to replace this with? I've had a look around the clubroom and the graveyard, but I don't really know what I'm looking for. [BOB] From president at ucc.asn.au Tue Feb 22 13:13:17 2011 From: president at ucc.asn.au (Matt Didcoe) Date: Tue, 22 Feb 2011 13:13:17 +0800 Subject: [tech] Old sun gear In-Reply-To: References:

Message-ID: Anyone want anything off this list? Contact Andrew directly. > On Sun, Feb 20, 2011 at 12:19 AM, Andrew Shackleton wrote: >> >> Oh yeah, theres a small amount of cisco gear. none available gb although it may be of use for someone building a prac lab. some bits do have a few gb "uplinks?" i forget what theyre called. maybe some WICS kicking around still and some old 16xx routers. >> >> On Sun, Feb 20, 2011 at 12:09 AM, Andrew Shackleton wrote: >>> >>> No delay at all. I left that for your logs as 5am was getting a bit too late to be looking for the UCC mailing list. >>> >>> I'm LLB at Murdoch and I was on UWA campus yesterday which reminded me of the UCC.. I think I was last up there 15 years ago. Dispense had jolt. >>> >>> Long story short this gear as I'm sure you know is heavy and there is little market for old enterprise technology about in Perth, most of it's not viable to even bother selling although it does pain me that it would just end up as unrecycled landfill. I would keep it although I just dont have the time for it anymore. >>> >>> I don't have a list of what is being disposed of although off the top of my head there are some e450's all fully loaded with spares.. one is brand new (despite being 10 years old) 220s, 280s etc (these crunch code fast, I think they are SpacrcII IIRC) , The quad cpu 4u versions of the 450's, ah there are at least 5 A1000 arrays and there should be a loaded fiber san t3 somewhere. >>> >>> I noticed reading your freshers guide that you have 2 x u80's.. I also have 2 of those kicking although I do recall ditching one of the ram expansion boards from one unit. They do however have some form of fancy video cards. >>> >>> There are also some sunrays, some still new in boxes for the avid collector if there is one amongst you.. I sadly dont have any cards kicking around for them for hot desking (which i find very cool, works over wan too although theyre $2 on ebay) >>> >>> Moving on from there there are some V2xx's v240s SparcIII dual cpu and a fully loaded v880, i think it has 64 gig ram. These units still have some commercial value and i'm tossing up what to do with them. There must be at least 50 spare scsi drives ranging from 18 - 72gb. >>> >>> Theres also some 1u pizza box p4's and a few 4u loaded dual xeons w scsi raid. now that i think of it theres some compaq dl 2u's although they are p3's.. and a dell 4u something.. >>> >>> I'm keen to join up with the club as Id like to have a tinker with some basic robotics and ill be on campus anyway. If the robogeeks want to work something out with the psychologically disturbed sun / unix crowd I'm more than happy to trade some of the decent gear. the rest of it is free. (and if all else fails you can have the lot.) >>> >>> I'm sure theres a lot more than that kicking around, i just havent been paying it much attention as of late. >>> >>> There are also some genuine sun racks. 42U, purple with dual power dist. They do look cool and i will miss them. >>> >>> As i remember there are also some d1000 arrays although i should get the data off of them first. maybe interested parties may like to give me a hand. ZFS as i'm sure youre familiar with is an incredible file system. >>> >>> Lots of scsi cables.. stacks of spares inc ram. >>> >>> Ill head down to the data centre tomorrow and see what else is kicking around. >>> >>> Apart from trade / barter for the better gear, this equipment is only being offered with the impression that members are not going to be taking it solely for resale or commercial use. That wouldnt be cool. Know what I mean ? >>> >>> Let me know if any of that sounds interesting. Its currently located in Melville and im even happy to help with moving some of it. I;ve got a truck and trailer. Get the freshers to carry it up the stairs into the ucc if any of its heading that way though please :) >>> >>> On a side note, there are also about 300 brand new CB style external speakers. someone must want a few of these. (Take them all, please, they are still neatly boxed in packs of 50 i think) Theres also a few m3 of esoteric electronics equipment. Ill have to sort it out to whats exactly in there. >>> >>> I knew I shouldn't have thrown out 100 mini lcd screens last week. I had just been tripping over them for a few months. Sorry. >>> >>> Oh, there's also 2 x 21" crt sun monitors. they came with the u80's and Ive been holding onto them for uber high res photo work although they can also go. >>> >>> If you want to touch base feel free to drop me a line : 0404 2100 40 >>> >>> More than happy to assist interested parties / UCC setup some of it as there are a few knacks and tricks that can really stump new players. >>> >>> Do you have a refurbish program for those that cant afford a pc ? There are some p4 desktops hanging around that need the boot. have to get the junior to nuke them first though. >>> >>> all gear generally has fcals and will run iSCSI if anyone wants to have a play with that.. >>> >>> Thats a bit of a jumbled email although i've been out in the sun all day and i only managed two hours of downtime last night. >>> >>> I dont have the time to be nuking all the data although ill be happy after a discussion with anyone interested to come to some arrangement for them to assist. I cant hand over half an old data centre of data without some assurances. A handshake will probably due with a threat of legal action ;) >>> >>> Thanks again for touching base. im up in subi if you want a coffee. What additional sun gear is the club interested in acquiring and ill see if its down there. >>> >>> On Sat, Feb 19, 2011 at 7:17 PM, Matt Didcoe wrote: >>>> >>>> Hey Andy, >>>> >>>> Sorry for the delay - just catching up with everything on IRC...was in >>>> at work this morning doing downtime! >>>> >>>> Do you have a list of what gear you have going? If you could send >>>> something through, I can bounce it around and see if there's anything >>>> the club or one of the members might like. >>>> >>>> Thanks, >>>> Matt >>>> >>>> -- >>>> Matt Didcoe >>>> President, University Computer Club >>>> president at ucc.asn.au >>> >> > From zanchey at ucc.gu.uwa.edu.au Mon Feb 28 13:08:52 2011 From: zanchey at ucc.gu.uwa.edu.au (David Adam) Date: Mon, 28 Feb 2011 13:08:52 +0800 (WST) Subject: [tech] Fwd: Fail event on /dev/md1:mylah Message-ID: Hooray, we dropped another disk on Mylah, and the RAID is now one disk away from world domination^W^Wtotal failure. (This RAID group hosts Mylah's /space, including /services and /away.) I'll back up the RAID to the SAN, but we really need to replace those failed drives or ditch the array. Mitch was looking at new drives on 2nd Feb, haven't heard anything since - perhaps we could fail over to the SAN in the interim. David Adam UCC Wheel Member zanchey at ucc.gu.uwa.edu.au ---------- Forwarded message ---------- Date: Mon, 28 Feb 2011 12:24:54 +0800 From: mdadm monitoring To: root at mylah.ucc.gu.uwa.edu.au Subject: Fail event on /dev/md1:mylah This is an automatically generated mail message from mdadm running on mylah A Fail event had been detected on md device /dev/md1. It could be related to component device /dev/sdg1. Faithfully yours, etc. P.S. The /proc/mdstat file currently contains the following: Personalities : [raid1] [raid6] [raid5] [raid4] md1 : active raid6 sdg1[8](F) sdn1[7] sdm1[6] sdl1[5] sdk1[4] sdi1[2] sdh1[1] 860231808 blocks level 6, 64k chunk, algorithm 2 [8/6] [_UU_UUUU] resync=DELAYED