From jmjoseph at andrew.cmu.edu Fri Dec 5 19:53:23 2003 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Fri, 05 Dec 2003 19:53:23 -0500 Subject: [auton-users] Liver Updates & SSH keys Message-ID: <3FD12883.5070804@andrew.cmu.edu> As Lazy seems to be working well for everyone, I've gone ahead with plans and made Liver identical to it. Please note that the ssh host key for Liver did change during the upgrade. You will likely need to remove the old ssh key from your ~/.ssh/known_hosts file. Simply search for "liver" in that file and remove the line. Next time you ssh, the new key will be added to the file. As always, be sure to let us know if you have any further difficulties. (admin at autonlab.org) -Jacob FYI, here is the public rsa key used on both liver and lazy: liver ssh # cat ssh_host_rsa_key.pub ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAIEAwVAH/U6iTsoQS8ysEgIbxWNj8i9a6xpfcai5DTVqW3xf Ltb8CiHB4PHfjFhfZJkZEk13nIO7spc74hv0tcyjAmcXRnu7zjLLpSdQk15TYI12UmlySw4O aYybYyLUDBIR2NGbYd/mLsj6rkCxWGDEyIwk6vnghHqLMRcrIzMVjOk= root at lazy.autonlab.org From jmjoseph at andrew.cmu.edu Mon Dec 8 02:15:05 2003 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Mon, 08 Dec 2003 02:15:05 -0500 Subject: [auton-users] Lira is down Message-ID: <3FD424F9.3040303@andrew.cmu.edu> Earlier today, Lira stopped responding and would not come back even after a reboot. After disassembly, it turns out that one of the CPUs decided to combust. (I am not joking--it is black in areas.) Anyway, until we get a replacement, I will put it back in action with 1 CPU, but in the morning. I need to grab a few parts from home first. Sorry for the inconvenience -Jacob From Andrew_W_Moore at lime.auton.cs.cmu.edu Mon Dec 8 08:56:25 2003 From: Andrew_W_Moore at lime.auton.cs.cmu.edu (Andrew W Moore) Date: Mon, 08 Dec 2003 08:56:25 -0500 Subject: [auton-users] Lira is down Message-ID: <20031208135645.A06D246@mail.autonlab.org> > After disassembly, it turns out that one of the CPUs decided to > combust. (I am not joking--it is black in areas.) Sounds like someone has a *really* bad bug in their code. From dpelleg+ at cs.cmu.edu Mon Dec 8 09:02:28 2003 From: dpelleg+ at cs.cmu.edu (Dan Pelleg) Date: Mon, 8 Dec 2003 09:02:28 -0500 Subject: [auton-users] Liver Updates & SSH keys In-Reply-To: <3FD12883.5070804@andrew.cmu.edu> References: <3FD12883.5070804@andrew.cmu.edu> Message-ID: <16340.33908.211866.856632@gs166.sp.cs.cmu.edu> Jacob Joseph writes: > As Lazy seems to be working well for everyone, I've gone ahead with > plans and made Liver identical to it. Please note that the ssh host key > for Liver did change during the upgrade. You will likely need to remove > the old ssh key from your ~/.ssh/known_hosts file. Simply search for > "liver" in that file and remove the line. Next time you ssh, the new > key will be added to the file. As always, be sure to let us know if you > have any further difficulties. (admin at autonlab.org) > > -Jacob > > > FYI, here is the public rsa key used on both liver and lazy: > Here is the fingerprint of that key. This is what SSH will show you when it's warning you about "authenticity of host can't be established". It should match the "RSA key fingerprint" SSH shows you. 46:3a:c7:58:10:49:bb:19:f9:c3:ef:fc:2f:09:79:71 Also note, that that existing key, which you should edit out, might also be in the file ~/.ssh/known_hosts2. From jmjoseph at andrew.cmu.edu Mon Dec 8 03:16:54 2003 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Mon, 08 Dec 2003 03:16:54 -0500 Subject: [auton-users] Loki intermission Message-ID: <3FD43376.20503@andrew.cmu.edu> Since no one is or has been using Loki today, and since I need to switch some cables, I am going to take it down for a few minutes right now. At worst, I don't expect more than a half hour. -Jacob From jmjoseph at andrew.cmu.edu Tue Dec 9 12:47:18 2003 From: jmjoseph at andrew.cmu.edu (Jacob Joseph) Date: Tue, 09 Dec 2003 12:47:18 -0500 Subject: [auton-users] Lina nfs problems Message-ID: <3FD60AA6.9020403@andrew.cmu.edu> It seems as though lina has been using the backup array we had in place when BigPapa was acting up. Apparently, the lofty was still accepting requests from it and forwarding it to the backup array even after BigPapa was put back in action. This has been fixed, but there is potential for data loss. If you wrote to "BigPapa" (your home directory, for example) from lina, that data will not be there now. It is not lost, but will need to be moved over. I can do this if you let me know what to merge. -Jacob From komarek at cmu.edu Wed Dec 17 16:09:54 2003 From: komarek at cmu.edu (Paul Komarek) Date: 17 Dec 2003 16:09:54 -0500 Subject: [auton-users] lira's replacement cpu Message-ID: <1071695394.3591.1.camel@laptop> Hello, We received a replacement cpu for lira today. Lira's first cpu is currently busy, and I will wait for that job to finish before installing the replacment for the failed cpu. Please do not start any new jobs on lira. -Paul Komarek From komarek at cmu.edu Wed Dec 17 18:11:09 2003 From: komarek at cmu.edu (Paul Komarek) Date: 17 Dec 2003 18:11:09 -0500 Subject: [auton-users] lira's replacement cpu In-Reply-To: <1071695394.3591.1.camel@laptop> References: <1071695394.3591.1.camel@laptop> Message-ID: <1071702669.3591.64.camel@laptop> On Wed, 2003-12-17 at 16:09, Paul Komarek wrote: > Hello, > > We received a replacement cpu for lira today. Lira's first cpu is > currently busy, and I will wait for that job to finish before installing > the replacment for the failed cpu. Please do not start any new jobs on > lira. > > -Paul Komarek > Lira's new cpu is installed and appears to be working normally. Feel free to use lira for two processes now. -Paul Komarek From pgunn at cs.cmu.edu Fri Dec 19 11:55:46 2003 From: pgunn at cs.cmu.edu (Pat Gunn) Date: Fri, 19 Dec 2003 11:55:46 -0500 (EST) Subject: [auton-users] Fedorization Message-ID: Hello all, Two things: While Artur is on vacation, I'm going to be upgrading his system to RedHat Fedora, so if you have any data on it or will be running any processes, know that there's a chance they'll be interrupted. I'll do my best to copy your data elsewhere and move it back afterwards, unless you tell me that it's not important to restore. The system name is logo. Secondly, if you're on vacation, and want your workstation upgraded to Fedora as well, let me know, along with notes on when you'll be gone and if you have any local data you want me to restore after the install. RedHat Fedora 1 is essentially RedHat Linux 10.0. Apart from the base install (which includes openoffice and other fun stuff), I'll also be adding OpenAFS (I can configure it to or not to use the AFS-style authentication) and anything else you might want (apart from Zephyr, which I currently don't know how to set up). In all fairness, here's a PRO/CON to moving your workstation to Fedora PRO ---------- It's much more current than Facilitized 7.1, and a bit more current than Facilitized 9 Facilities won't be messing with your system and breaking things You won't need to deal with existing Facilities breakage It's easier to install additional software (there's a tool called yum, and also if you need to compile things, facilities won't erase it automagically) Hardware support's a little bit better If you need to play windows/mac media or DVDs, I can install a package called mplayer for you. CON ---------- No more Zephyr. A lot of us don't use it, but if you do, this is a minus Facilities won't support your software (I'll be happy to) Setting up some site-licensed software like Matlab might be more involved (read: I haven't dealt with the paperwork involved, if any, so no idea) No kinit, so you'll need to klog occasionally. Note that this is also the case with Facilitized 9 -- Pat Gunn Research/Systems Programmer, Auton Group, CMU From pgunn at cs.cmu.edu Tue Dec 23 12:28:28 2003 From: pgunn at cs.cmu.edu (Pat Gunn) Date: Tue, 23 Dec 2003 12:28:28 -0500 (EST) Subject: [auton-users] RSA key for logo Message-ID: Hey all, 1024 8c:c2:a9:fa:20:29:a2:a2:46:e3:77:ef:99:35:73:e1 /etc/ssh/ssh_host_rsa_key.pub is the new rsa fingerprint for logo. It's running Fedora1 now. A few of you had data on its local disks. I saved tarballs, and can restore them if you like (check /home for them). I enabled hyperthreading (the kernel supports it and should do reasonable things, scheduler-wise), so it'll look like it has 4 CPUs. Hyperthreading splits each CPU into two 'virtual CPUs', using the pipelining and some seperate hardware on each CPU to do the trick. As some parts are shared, it's not as good as SMP, but it is, with the right scheduler (which the Fedora kernels have, I believe), better than trying to run the same set of jobs with Hyperthreading disabled. Let me know if anything isn't as it should be, or if you want anything installed on it. It, of course, still talks NFS and AFS. -- Pat Gunn Research/Systems Programmer, Auton Group, CMU