From Andrew_W_Moore at lime.auton.cs.cmu.edu Fri Jan 2 16:22:21 2004 From: Andrew_W_Moore at lime.auton.cs.cmu.edu (Andrew W Moore) Date: Fri, 02 Jan 2004 16:22:21 -0500 Subject: [auton-users] Lab network connections down Message-ID: <20040102215045.5855C18E85A@mx2.zoneedit.com> The whole of the "A" floor of Newell-Simon has a network outage (the other labs there are experiencing the same problem) so sadly no use of the lab number crunchers, web page or file server for the time being. Andrew From jmjoseph at andrew.cmu.edu Sat Jan 3 02:41:20 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Sat, 03 Jan 2004 02:41:20 -0500 Subject: [auton-users] Lab is back on Message-ID: <3FF67220.2010109@andrew.cmu.edu> I'm happy to announce that the AUTON lab is operational again. Thanks to Pat, the necessary power buttons were pushed and all machines residing in the lab are back up. Please let us know at admin at autonlab.org with any questions or problems. -Jacob From komarek at cmu.edu Sat Jan 10 10:37:36 2004 From: komarek at cmu.edu (Paul Komarek) Date: 10 Jan 2004 10:37:36 -0500 Subject: [auton-users] lina is down Message-ID: <1073749056.21880.4.camel@laptop> Hi everyone, It looks like the lopX servers arrived just in time. Lina is a twin of lira, and lina just had the same cpu burn out that lira did (the second cpu). I think lina will run on the remaining cpu at this point, but I'm not sure I want to bother. If someone wants/needs lina's remaining cpu running, please email admin at autonlab.org. We'll look into replacement, but it could be tricky. To see why, you can browse the photos at lazy:/tmp/lina_cpu2. The motherboard appears to be damaged, and the vendor of these machines is one we have quit using due to service problems. Thanks goes to the Nourbakhsh lab for loan of the digicam. -Paul Komarek From dpelleg+ at cs.cmu.edu Wed Jan 7 15:27:22 2004 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Wed, 7 Jan 2004 15:27:22 -0500 Subject: [auton-users] lina is down In-Reply-To: <1073749056.21880.4.camel@laptop> References: <1073749056.21880.4.camel@laptop> Message-ID: <16380.27562.782201.425388@localhost.localdomain> I placed a copy of the photos, with thumbnails and smaller-size options, at: http://web.cs.cmu.edu/~dpelleg/cpu/ From komarek at cmu.edu Thu Jan 8 18:41:58 2004 From: komarek at cmu.edu (Paul Komarek) Date: Thu, 08 Jan 2004 18:41:58 -0500 Subject: [auton-users] No AFS on lop* yet Message-ID: <1073605317.31120.17.camel@localhost> Hi everyone, Andrew had me say a short blurb about the new lop* servers at the lab meeting yesterday. I said they were up and ready. There's a caveat, though. AFS is not yet installed. If you want to work on a lop* and need AFS for cvs access but no other reason, then I (Paul) can help you quickly transition to cvs over ssh. I'll bet Pat (sorry to volunteer you, Pat) can help, too. I believe Jacob has successfully tested AFS on lop5, but has not finished full installation on any of the lop*. -Paul Komarek From komarek at cmu.edu Fri Jan 9 13:30:50 2004 From: komarek at cmu.edu (Paul Komarek) Date: Fri, 09 Jan 2004 13:30:50 -0500 Subject: [auton-users] [Fwd: Power Off Saturday Morning in NSH, PRB and Smith] Message-ID: <1073673049.1002.16.camel@localhost> Hello everyone, Remember that power in several buildings will be off between midnight and 8am Saturday morning. It is unlikely that your UPS will survive that long, so you may want to shutdown your computers tonight. Remember that others may be using your computer. Before you shut down, you might want to run "top" to see if there are active jobs (and get the username). For those of us running jobs on other people's desktops, I recommend terminating those jobs by 4pm today. We haven't decided yet when to shut down the lab computers. Please send email to admin at autonlab.org with information about your expected usage of lab computers tonight. However, it's unlikely any admin will volunteer to wait until 11:59pm to shut down the lab. We'll post a lab shutdown time once one has been selected. -Paul -----Forwarded Message----- From: Jim Skees To: ri-edsh at cs.cmu.edu, ri-nsh at cs.cmu.edu, rjg+ at andrew.cmu.edu, lenscmu at cs.cmu.edu, cs-students at burks-pc.adm.cs.cmu.edu, cs-faculty at burks-pc.adm.cs.cmu.edu, cs-technical at burks-pc.adm.cs.cmu.edu, cs-support-staff at burks-pc.adm.cs.cmu.edu, cs-visitors at burks-pc.adm.cs.cmu.edu, hcii-members at cs.cmu.edu, lti-faculty-core at cs.cmu.edu, lti-students at cs.cmu.edu, lti-staff at cs.cmu.edu, isri-people at cs.cmu.edu, arif+ at andrew.cmu.edu, tcarr at cs.cmu.edu, aw16+ at andrew.cmu.edu, cleah+ at cs.cmu.edu, cald-people at cs.cmu.edu, sw2x+ at andrew.cmu.edu, pcampbel at cs.cmu.edu, hollinge+ at cs.cmu.edu, mkrabin+ at andrew.cmu.edu, huyh+ at andrew.cmu.edu, paul.komarek at cs.cmu.edu Subject: Power Off Saturday Morning in NSH, PRB and Smith Date: Tue, 06 Jan 2004 15:26:00 -0500 Folks, All power to Newell-Simon, Smith, and Planetary Robotics will be off from midnight Friday until 8 a.m. Saturday (i.e. Jan. 9-10). This shutdown will permit FMS to replace the main switch that controls power to these and other buildings in the former Bureau of Mines complex. SCS folks who use offices and labs in these buildings should make sure their computing equipment is properly shut down and powered off before they leave for the day Friday. The eight-hour length of the shutdown is quite a bit longer than your UPS units can handle. If you have any questions or concerns, please send e-mail to skees at cs. --Jim >Division of Administration >Facilities Management Services > >TO: Hamburg Hall > Smith Hall > Publications and Printing > 4902 Forbes Ave > Planetary Robotics Building > Newell-Simone Hall > Shutdown Group >FROM: Service Response Center >DATE: January 6,2004 > >SUBJECT: Electrical Shutdown, Saturday, January 10 > >All electric service to Hamburg Hall, Smith Hall, Publications and >Printing, Planetary Robotics, Newell-Simon and 4902 Forbes Ave (Student >Center) will be interrupted from 12:01 AM to 8:00 AM on Saturday, January >10, 2004. The reason for this shutdown is to repair the main disconnect >switch that connects these buildings to the campus electrical grid. > >For further information, contact Kyle Tomer at 8-6332 (email >kt2n at andrew.cmu.edu) James J. Skees Building Facilities Manager School of Computer Science Carnegie Mellon University 5000 Forbes Ave. Pittsburgh, PA 15213 Office: Wean Hall 4210 (412) 268-7653 voice (412) 268-5576 fax From komarek at cmu.edu Fri Jan 9 17:38:45 2004 From: komarek at cmu.edu (Paul Komarek) Date: Fri, 09 Jan 2004 17:38:45 -0500 Subject: [auton-users] lab shutdown at 11pm Message-ID: <1073687925.1009.126.camel@localhost> Hello everyone, Pat has kindly offered to shut down the lab sometime after 11pm. Please finish all work before then. At some point Pat will become a merciless savage and *terminate* your jobs after 11pm; and all of the admins will support him! And if he doesn't, the power gods will certainly have the final word. I will bring the lab up by 11am tomorrow morning. If you want the lab up before then, please let me know. Remember that the power is not guaranteed to be restored until 8am (and they didn't use the word guarantee for 8am, either). -Paul Komarek From komarek at cmu.edu Fri Jan 9 18:52:31 2004 From: komarek at cmu.edu (Paul Komarek) Date: Fri, 09 Jan 2004 18:52:31 -0500 Subject: [auton-users] lab temps and early shutdown Message-ID: <1073692351.1008.143.camel@localhost> Hello everyone, Lab temperatures are soaring. I'm guessing they shutdown the air conditioning early. I'm going to turn off all machines that aren't running jobs from home. I'll shut down the remaining machines w/in an hour, I hope. -Paul From komarek at cmu.edu Fri Jan 9 19:05:53 2004 From: komarek at cmu.edu (Paul Komarek) Date: 09 Jan 2004 19:05:53 -0500 Subject: [auton-users] lab temps and early shutdown In-Reply-To: <1073692351.1008.143.camel@localhost> References: <1073692351.1008.143.camel@localhost> Message-ID: <1073693153.23138.0.camel@laptop> Due to 44C+ temps on critical infrastructure in the lab, I'm going to continue shutting down all services including the web server. -Paul On Fri, 2004-01-09 at 18:52, Paul Komarek wrote: > Hello everyone, > > Lab temperatures are soaring. I'm guessing they shutdown the air > conditioning early. I'm going to turn off all machines that aren't > running jobs from home. I'll shut down the remaining machines w/in an > hour, I hope. > > -Paul > From komarek at cmu.edu Sat Jan 10 11:46:58 2004 From: komarek at cmu.edu (Paul Komarek) Date: 10 Jan 2004 11:46:58 -0500 Subject: [auton-users] lab temps still high Message-ID: <1073753218.25373.11.camel@laptop> Hello, The lab temps are still around 78F with no machines running. I have brought the file server and web server up, but we cannot start any compute machines (lop*, alphas, lira) until we have climate control. -Paul From komarek at cmu.edu Sat Jan 10 14:51:41 2004 From: komarek at cmu.edu (Paul Komarek) Date: 10 Jan 2004 14:51:41 -0500 Subject: [auton-users] lab machines are up! Message-ID: <1073764301.28457.6.camel@laptop> Hello, I'm happy to announce that the lab machines are up. The temp is holding at 72F, which is the best we can hope for given the AC problems related to the NSH B level flood. AFS is not yet available on all of the alphas due to a kernel upgrade (for security reasons), but will be later today. AFS on lop* should be available by next weekend. As you might have noticed, keeping AFS running is somewhat painful and uses much of our admin time. -Paul Komarek From komarek at cmu.edu Sat Jan 10 14:53:11 2004 From: komarek at cmu.edu (Paul Komarek) Date: 10 Jan 2004 14:53:11 -0500 Subject: [auton-users] AFS is up on alphas Message-ID: <1073764390.28457.9.camel@laptop> Hi, I accidently sent that last email before I had a chance to delete my (stale) warning about AFS on the Alphas. With Jacob's help via mobile phones, AFS is now running on the Alphas. -Paul From komarek at cmu.edu Sat Jan 10 15:52:27 2004 From: komarek at cmu.edu (Paul Komarek) Date: 10 Jan 2004 15:52:27 -0500 Subject: [auton-users] more temp problems Message-ID: <1073767946.2914.5.camel@laptop> Hi, After letting all the machines run for a bit, our temp climbed back to 78F. This is too close to the "scary" region for comfort. Since I don't feel like babysitting the lab all day, I've shutdown our hottest machines and the least essential machines. Available: lop* loki Unavailable: lazy liver lira lina (but for other reasons =-) loon There was a massive air conditioning failure in NSH that won't be fixed until next week some time. We'll keep lop* and loki running so long as the lab temps stay low enough. Though unintuitive, more cooling capacity will become available if the outside air temp rises. -Paul From komarek at cmu.edu Mon Jan 12 13:30:52 2004 From: komarek at cmu.edu (Paul Komarek) Date: Mon, 12 Jan 2004 13:30:52 -0500 Subject: [auton-users] AFS on lop* Message-ID: <1073932251.1008.188.camel@localhost> Hello everyone, I have been reanswering questions about AFS on lop* this morning. While AFS is not yet available, it will be will be installed. Jacob returned from holiday yesterday evening and will take care of it as soon as possible. He as already successfully tested AFS on lop*, but has not had a chance to "productize" AFS for the lop* machines. If lack of AFS for CVS is preventing you from working on the lop*, you have three options until AFS is installed: 1) Do you normal CVS/AFS work on loki. Once the air conditioning is fixed, you will be able to use lira, lazy and liver for the same purpose. Of course, you should recompile on the lop* once you have your CVS tree updated. 2) Use CVS over ssh instead. This is what we have to do at home to access ssh (unless you run an AFS client and krb stack at home). It is also the standard standard way of running secure CVS outside of CMU. I will be happy to help anyone move to ssh from AFS for CVS. 3) Wait for Jacob to install AFS. This message is (more-or-less) a duplicate of the one sent on 8 January, and can be safely ignored. -Paul Komarek From dpelleg+ at cs.cmu.edu Mon Jan 12 13:44:19 2004 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 12 Jan 2004 13:44:19 -0500 Subject: [auton-users] AFS on lop* In-Reply-To: <1073932251.1008.188.camel@localhost> References: <1073932251.1008.188.camel@localhost> Message-ID: <16386.60163.918128.953630@gs166.sp.cs.cmu.edu> Paul Komarek writes: > Hello everyone, > > If lack of AFS for CVS is preventing you from working on the lop*, you > have three options until AFS is installed: > > 1) Do you normal CVS/AFS work on loki. Once the air conditioning is > fixed, you will be able to use lira, lazy and liver for the same > purpose. Of course, you should recompile on the lop* once you have your > CVS tree updated. Alternatively, 1a. Do your normal CVS/AFS work on your desktop machine, or any machine which has access to the BigPapa volume over NFS (all of the "desktop servers" - loco, logo etc. - do). Simply cd to the appropriate directory (something like /mnt/BigPapa/home/USER/h/PROJECT ) and then do the cvs operation. This is the same directory as h/PROJECT under your home directory on the lab machines. > > 2) Use CVS over ssh instead. This is what we have to do at home to > access ssh (unless you run an AFS client and krb stack at home). It is > also the standard standard way of running secure CVS outside of CMU. I > will be happy to help anyone move to ssh from AFS for CVS. Before you bug Paul, We have a guide for this (from the "Programming" section of the "old site" on www.autonlab.org): http://www.autonlab.org/doc/cvs.html#4 I'm not sure it's still current. I am sure I did not get any bug reports about it :) From komarek at cmu.edu Wed Jan 14 15:16:33 2004 From: komarek at cmu.edu (Paul Komarek) Date: Wed, 14 Jan 2004 15:16:33 -0500 Subject: [auton-users] reserving lop2 Message-ID: <1074111393.14106.5.camel@localhost> Hello everyone, I'd like to request exclusive access to lop2 for one week (until Wednesday Jan 21). I have a queue of several hundred experiments to run for my thesis, and would like to keep them on one of the lop* since the previous 2500+ experiments were also run on lop*. Does anyone have an objection to this? I can start the reservation any time. I suppose I need to propose a real time, though, so how about tonight at midnight? -Paul From komarek at cmu.edu Wed Jan 14 17:32:32 2004 From: komarek at cmu.edu (Paul Komarek) Date: Wed, 14 Jan 2004 17:32:32 -0500 Subject: [auton-users] reserving lop2 In-Reply-To: <1074111393.14106.5.camel@localhost> References: <1074111393.14106.5.camel@localhost> Message-ID: <1074119551.14279.10.camel@localhost> Now that lop2 is no longer busy, I have made the reservation. -Paul Komarek On Wed, 2004-01-14 at 15:16, Paul Komarek wrote: > Hello everyone, > > I'd like to request exclusive access to lop2 for one week (until > Wednesday Jan 21). I have a queue of several hundred experiments to run > for my thesis, and would like to keep them on one of the lop* since the > previous 2500+ experiments were also run on lop*. > > Does anyone have an objection to this? I can start the reservation any > time. I suppose I need to propose a real time, though, so how about > tonight at midnight? > > -Paul > From jmjoseph at andrew.cmu.edu Fri Jan 16 23:20:26 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Fri, 16 Jan 2004 23:20:26 -0500 Subject: [auton-users] New status graphs Message-ID: <4008B80A.6010807@andrew.cmu.edu> We're happy to announce that the new Auton Status backend is up and running. Information of all of the lops has been added, along with a new "Server" page to see the status of our backend servers. We have selected colors with higher contrast but are happy to change them if someone can suggest 10-15 more distinct colors(perferably in hex). We believe these are a great improvement. Let us know what you think. It's all at: http://www.autonlab.org/status -Jacob and Paul From komarek at cmu.edu Sat Jan 17 14:53:59 2004 From: komarek at cmu.edu (Paul Komarek) Date: Sat, 17 Jan 2004 14:53:59 -0500 Subject: [auton-users] New status graphs In-Reply-To: <4008B80A.6010807@andrew.cmu.edu> References: <4008B80A.6010807@andrew.cmu.edu> Message-ID: <1074369239.10229.7.camel@localhost> One thing to note: we haven't been collecting data on the lop* until very recently, with the exception of lop5 for which data has been collected slightly longer. -Paul On Fri, 2004-01-16 at 23:20, Jacob Joseph wrote: > We're happy to announce that the new Auton Status backend is up and > running. Information of all of the lops has been added, along with a > new "Server" page to see the status of our backend servers. > > We have selected colors with higher contrast but are happy to change > them if someone can suggest 10-15 more distinct colors(perferably in > hex). We believe these are a great improvement. Let us know what you > think. > > It's all at: http://www.autonlab.org/status > > > -Jacob and Paul > > From jmjoseph at andrew.cmu.edu Tue Jan 20 03:50:23 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Tue, 20 Jan 2004 03:50:23 -0500 Subject: [auton-users] AFS on lop* Message-ID: <400CEBCF.2040201@andrew.cmu.edu> After discovering a major bug with respect to failed writes to AFS on the Opterons, I have fixed the afflicted machines and taken the opportunity to install on the rest. AFS is now functional on ALL of the lops. Please let me know if you notice anything out of the ordinary. Thanks, -Jacob From dpelleg+ at cs.cmu.edu Tue Jan 20 10:10:45 2004 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Tue, 20 Jan 2004 10:10:45 -0500 Subject: [auton-users] AFS on lop* In-Reply-To: <400CEBCF.2040201@andrew.cmu.edu> References: <400CEBCF.2040201@andrew.cmu.edu> Message-ID: <16397.17653.497308.245204@localhost.localdomain> Jacob Joseph writes: > After discovering a major bug with respect to failed writes to AFS on > the Opterons, I have fixed the afflicted machines and taken the > opportunity to install on the rest. AFS is now functional on ALL of the > lops. Please let me know if you notice anything out of the ordinary. > > Thanks, > -Jacob > I seem to have lost my kerberos tokens in the middle of a cvs update. I logged in and did a klog. Then I did a "cvs up" in one directory, which went well. Then in another directory. A few files updated ok, then I got this: cvs [update aborted]: cannot close history file: /mnt/BigPapa/home/dpelleg/g/cvsroot/CVSROOT/history: Permission denied cvs update: failed to remove lock /afs/cs/project/learn/group/cvsroot/dset/#cvs.rfl.lop2.autonlab.org.12400: Permission denied I wasn't able to fs la that directory from lop2 ("permission denied"). I was, however, able to do so from another (facilitized) machine holding the same credentials. I did another "klog" on lop2 and then another cvs up, which succeeded. Looks like a nasty cache or kerberos problem; in the meantime, if you get this, you should try again after a "klog". From jmjoseph at andrew.cmu.edu Tue Jan 20 12:03:04 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Tue, 20 Jan 2004 12:03:04 -0500 Subject: [auton-users] AFS on lop* In-Reply-To: <16397.17653.497308.245204@localhost.localdomain> References: <400CEBCF.2040201@andrew.cmu.edu> <16397.17653.497308.245204@localhost.localdomain> Message-ID: <400D5F48.1070000@andrew.cmu.edu> Dan Pelleg wrote: > Jacob Joseph writes: > > After discovering a major bug with respect to failed writes to AFS on > > the Opterons, I have fixed the afflicted machines and taken the > > opportunity to install on the rest. AFS is now functional on ALL of the > > lops. Please let me know if you notice anything out of the ordinary. > > > > Thanks, > > -Jacob > > > > I seem to have lost my kerberos tokens in the middle of a cvs update. I > logged in and did a klog. Then I did a "cvs up" in one directory, which > went well. Then in another directory. A few files updated ok, then I got > this: > > cvs [update aborted]: cannot close history file: /mnt/BigPapa/home/dpelleg/g/cvsroot/CVSROOT/history: Permission denied > cvs update: failed to remove lock /afs/cs/project/learn/group/cvsroot/dset/#cvs.rfl.lop2.autonlab.org.12400: Permission denied > > > I wasn't able to fs la that directory from lop2 ("permission denied"). I > was, however, able to do so from another (facilitized) machine holding the > same credentials. > > I did another "klog" on lop2 and then another cvs up, which succeeded. > > Looks like a nasty cache or kerberos problem; in the meantime, if you get > this, you should try again after a "klog". Dan, this is the "major bug". Paul has reserved lop2, preventing user access, and no-one but us admins has access. It will be fixed like the others as soon as Paul is done. He does not need AFS in the meantime. -Jacob From komarek at cmu.edu Tue Jan 20 12:03:08 2004 From: komarek at cmu.edu (Paul Komarek) Date: Tue, 20 Jan 2004 12:03:08 -0500 Subject: [auton-users] AFS on lop* In-Reply-To: <16397.17653.497308.245204@localhost.localdomain> References: <400CEBCF.2040201@andrew.cmu.edu> <16397.17653.497308.245204@localhost.localdomain> Message-ID: <1074618187.10226.119.camel@localhost> On Tue, 2004-01-20 at 10:10, Dan Pelleg wrote: > Jacob Joseph writes: > > After discovering a major bug with respect to failed writes to AFS on > > the Opterons, I have fixed the afflicted machines and taken the > > opportunity to install on the rest. AFS is now functional on ALL of the > > lops. Please let me know if you notice anything out of the ordinary. > > > > Thanks, > > -Jacob > > > > I seem to have lost my kerberos tokens in the middle of a cvs update. I > logged in and did a klog. Then I did a "cvs up" in one directory, which > went well. Then in another directory. A few files updated ok, then I got > this: > > cvs [update aborted]: cannot close history file: /mnt/BigPapa/home/dpelleg/g/cvsroot/CVSROOT/history: Permission denied > cvs update: failed to remove lock /afs/cs/project/learn/group/cvsroot/dset/#cvs.rfl.lop2.autonlab.org.12400: Permission denied > > > I wasn't able to fs la that directory from lop2 ("permission denied"). I > was, however, able to do so from another (facilitized) machine holding the > same credentials. > > I did another "klog" on lop2 and then another cvs up, which succeeded. > > Looks like a nasty cache or kerberos problem; in the meantime, if you get > this, you should try again after a "klog". > Hi Dan, lop2 is reserved for me right now, and I asked Jacob not to do any AFS stuff on it in case of a kernel problem. lop2 will be updated when my reservation has lapsed. Because you are an admin, you can still login. Nobody else can, however. -Paul From Weng-Keen_Wong at gs54.sp.cs.cmu.edu Wed Jan 21 12:37:26 2004 From: Weng-Keen_Wong at gs54.sp.cs.cmu.edu (Weng-Keen Wong) Date: Wed, 21 Jan 2004 12:37:26 -0500 Subject: [auton-users] Thesis Defense Message-ID: <20040121173803.907073D0@mail.autonlab.org> Hi everyone, Just letting you know that my thesis defense is this Thurs (Jan 22nd) 10:30 at Wean 4623. If you're awake at that time, feel free to attend =) Weng-Keen From anya at cmu.edu Wed Jan 21 14:22:32 2004 From: anya at cmu.edu (Anna Goldenberg) Date: Wed, 21 Jan 2004 14:22:32 -0500 Subject: [auton-users] reservation of lop3 Message-ID: <017101c3e053$eb1ab7d0$c600a8c0@Anna> hi, i would like to reserve lop3 at least till the end of the week to finish running experiments for my icml paper. if you have any objections, please let me know asap, otherwise it'll be reserved by 4:30pm this afternoon. thank you! anna -------------- next part -------------- An HTML attachment was scrubbed... URL: From komarek at cmu.edu Wed Jan 21 17:09:40 2004 From: komarek at cmu.edu (Paul Komarek) Date: Wed, 21 Jan 2004 17:09:40 -0500 Subject: [auton-users] reservation of lop3 In-Reply-To: <017101c3e053$eb1ab7d0$c600a8c0@Anna> References: <017101c3e053$eb1ab7d0$c600a8c0@Anna> Message-ID: <1074722979.10229.232.camel@localhost> I have reserved lop3 for Anya until midnight on Saturday the 24th of January (since she didn't really give an end date). -Paul On Wed, 2004-01-21 at 14:22, Anna Goldenberg wrote: > hi, i would like to reserve lop3 at least till the end of the week to > finish running experiments for my icml paper. if you have any > objections, please let me know asap, otherwise it'll be reserved by > 4:30pm this afternoon. > > thank you! > anna From jmjoseph at andrew.cmu.edu Wed Jan 21 20:21:17 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Wed, 21 Jan 2004 20:21:17 -0500 Subject: [auton-users] de-reservation of lop2 In-Reply-To: <1074722979.10229.232.camel@localhost> References: <017101c3e053$eb1ab7d0$c600a8c0@Anna> <1074722979.10229.232.camel@localhost> Message-ID: <400F258D.8040303@andrew.cmu.edu> Don't know if de-reservation is valid english, but Lop2 is now available for general usage. -Jacob From komarek at cmu.edu Thu Jan 22 17:23:47 2004 From: komarek at cmu.edu (Paul Komarek) Date: Thu, 22 Jan 2004 17:23:47 -0500 Subject: [auton-users] lop5 kernel change Message-ID: <1074810227.31412.37.camel@localhost> Hello everyone, I have changed the kernel of lop5 to version 2.6.1-mm4 for general speed and stability testing. If you have a chance, please run your code on lop5 and compare it to previous results you might have (or to results on one of the other lops). If we encounter any stability problems with 2.6.1-mm4, we will revert to 2.4 immediately. If 2.6.1-mm4 works well for everyone, we will eventually revert to 2.4 to restore AFS functionality (see note below), but will use the performance data to help with future decisions. lop5 will return to kernel 2.4 before Monday, if not sooner. Please note that AFS is not yet available for kernel 2.6.1-mm4 (for any platform). It is possible that AFS will never be available for kernels beyond 2.4, given the current lack of enthusiasm for fixing openafs on the openafs mailing list. If this is the case, we will be stuck with 2.4 forever on all machines that require afs. -Paul From komarek at cmu.edu Thu Jan 22 17:36:12 2004 From: komarek at cmu.edu (Paul Komarek) Date: Thu, 22 Jan 2004 17:36:12 -0500 Subject: [auton-users] lop5 kernel change In-Reply-To: <1074810227.31412.37.camel@localhost> References: <1074810227.31412.37.camel@localhost> Message-ID: <1074810971.31413.39.camel@localhost> I should mention that I'm happy to move lop5 back to 2.4 if desired. -Paul On Thu, 2004-01-22 at 17:23, Paul Komarek wrote: > Hello everyone, > > I have changed the kernel of lop5 to version 2.6.1-mm4 for general speed > and stability testing. If you have a chance, please run your code on > lop5 and compare it to previous results you might have (or to results on > one of the other lops). > > If we encounter any stability problems with 2.6.1-mm4, we will revert to > 2.4 immediately. If 2.6.1-mm4 works well for everyone, we will > eventually revert to 2.4 to restore AFS functionality (see note below), > but will use the performance data to help with future decisions. lop5 > will return to kernel 2.4 before Monday, if not sooner. > > Please note that AFS is not yet available for kernel 2.6.1-mm4 (for any > platform). It is possible that AFS will never be available for kernels > beyond 2.4, given the current lack of enthusiasm for fixing openafs on > the openafs mailing list. If this is the case, we will be stuck with > 2.4 forever on all machines that require afs. > > -Paul > From komarek at cmu.edu Fri Jan 23 03:26:20 2004 From: komarek at cmu.edu (Paul Komarek) Date: Fri, 23 Jan 2004 03:26:20 -0500 Subject: [auton-users] lop5 back to kernel 2.4 with AFS Message-ID: <1074846370.31414.81.camel@localhost> lop5 is back to kernel 2.4 with AFS. Which means it's just like the other lops again. -Paul komarek From sabhnani at andrew.cmu.edu Fri Jan 23 16:08:43 2004 From: sabhnani at andrew.cmu.edu (Maheshkumar Sabhnani) Date: Fri, 23 Jan 2004 16:08:43 -0500 Subject: [auton-users] Info. Message-ID: <40118D5B.4010907@andrew.cmu.edu> Hello everyone, If your project uses frequent incremental datset building like appfbrun does, you might now experience 2-3 times speed ups in your program execution. Don't be surprised, just thank ANDREW. He found some optimization issues in the code that required a couple of additions/modifications to files dset/pivec.c and dset/column.c. I have done some NAIVE speed comparisons. I noticed 2-15 fold speed ups for datset processing. In case you find any bugs, please feel free send me comments. If not, then enjoy faster running programs !!!! Robin From komarek at cmu.edu Sat Jan 24 21:55:14 2004 From: komarek at cmu.edu (Paul Komarek) Date: Sat, 24 Jan 2004 21:55:14 -0500 Subject: [auton-users] lazy, liver and loon are no longer in service Message-ID: <1074999314.8817.10.camel@localhost> We have decomissioned lazy (Alpha), liver (Alpha) and loon (P-II). If you previously had an account only one of these machines, your account will now work on loki.autonlab.org. Please let us know if you have any troubles with the new arrangements. If you need access to a decomissioned machine, contact us. If you want to make an offer to purchase a decomissioned machine, meet me by the lab -- cash only, please ;-) -Paul Komarek From jmjoseph at andrew.cmu.edu Mon Jan 26 17:26:21 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Mon, 26 Jan 2004 17:26:21 -0500 Subject: [auton-users] LINA is back Message-ID: <4015940D.8@andrew.cmu.edu> Hi. I've replaced the burned parts in lina and put it back in service. As always, let me know if there are any problems. Happy computing. -Jacob From jmjoseph at andrew.cmu.edu Wed Jan 28 12:52:57 2004 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Wed, 28 Jan 2004 12:52:57 -0500 Subject: [auton-users] LAB OUTAGE Message-ID: <4017F6F9.1030909@andrew.cmu.edu> Jim Skees has just let me know that we're going to be without air conditioning this Saturday, the 31st, from 7am-11pm. I know some of you are still processing for the ICML deadline in particular, but this is going to necessitate turning off most if not all of the compute machines from Friday night to Sunday morning. I will of course leave busy machines to the end, but if termperatures get too high, I will have to shut everything down. So, starting any long-running jobs now is probably futile. If this is going to cause you any major problems, let me know. It may be possible to provide enough cooling some other way if we have some advanced notice. -Jacob From komarek at cmu.edu Wed Jan 28 14:49:05 2004 From: komarek at cmu.edu (Paul Komarek) Date: Wed, 28 Jan 2004 14:49:05 -0500 Subject: [auton-users] Reserving lop2 Message-ID: <1075319345.23855.20.camel@localhost> Hello, I would like to reserve lop2 (again) for one week, starting tonight around 5pm. Again this is for my thesis, for time-sensitive experiments. Are there any objections? -Paul Komarek From anya at cs.cmu.edu Wed Jan 28 15:07:48 2004 From: anya at cs.cmu.edu (Anna Goldenberg) Date: Wed, 28 Jan 2004 15:07:48 -0500 (EST) Subject: [auton-users] reservation of lop3 Message-ID: Hi! I would like to appologize for extending the reservation of lop3. The process should have been done a couple of days ago, but it is not (which is good for us, because it's a competitor algorithm :) Anyway, I would like to ask for permission to keep the reservation of lop3 until Tuesday of next week (the deadline for icml), unless there are any objections. I am happy to give up the reservation, but I would rather not kill the process as it seems to be almost done and I do have a couple of other experiments to run for icml. Thank you and again my appologies for extending the reservation Anna From pgunn at cs.cmu.edu Thu Jan 29 11:25:47 2004 From: pgunn at cs.cmu.edu (Pat Gunn) Date: Thu, 29 Jan 2004 11:25:47 -0500 (EST) Subject: [auton-users] Project sprat_pgi retired (NOT sprat_pgi3) Message-ID: Hey all, The project sprat_pgi has been retired, and moved to the retired version of the CVS repository. If you see any errors about trying to update this project (e.g. you do a "cvs update" in your h directory), remove the sprat_pgi directory and all will be good. If you think it was wrong to retire sprat_pgi, let me know, and I'll move it back. As of present there are two other versions of sprat that are alive and kicking, "sprat" and "sprat_pgi3" (although ideally they'll eventually merge back together). -- Pat Gunn Research/Systems Programmer, Auton Group, CMU From sabhnani at cs.cmu.edu Thu Jan 29 11:51:39 2004 From: sabhnani at cs.cmu.edu (Maheshkumar Sabhnani) Date: Thu, 29 Jan 2004 11:51:39 -0500 Subject: [auton-users] Andrew's old laptop. Message-ID: <40193A1B.5090404@cs.cmu.edu> Hello everyone, I am not sure if I mentioned this before. Anyways I am currently using Andrew's old laptop. It has both windows and Linux installed on it. In case anyone wants to use it for a few days, (like for conferences, presentations, or research work) feel free to send me an email. Robin. From komarek at cmu.edu Fri Jan 30 03:13:48 2004 From: komarek at cmu.edu (Paul Komarek) Date: Fri, 30 Jan 2004 03:13:48 -0500 Subject: [auton-users] Air Conditioning outage on Saturday Message-ID: <1075450427.23855.165.camel@localhost> Hello, This email is a follow-up to Jacob's LAB OUTAGE email to users at autonlab.org on Wed 28 Jan. There will be an air conditioning outage on Saturday 31 January, all day long. During this outage the problem that created the previous air conditioning outage will be outed and a soluton worked out. The outcome: we cannot keep all of our machines running on Saturday. We intend to power off all the cycle servers all of Saturday, and the file server and web server will be offline for some time as well. If this puts anyone off, please email admin at autonlab.org with your conerns. We understand their is an ICML deadline fast approaching, and may be able to arrange partial service for some cycle servers to offset your losses. The web and file server outages are for upgrades and should be short-lived. -Paul Komarek