From predragp at andrew.cmu.edu Thu Oct 2 13:30:10 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Thu, 2 Oct 2014 13:30:10 -0400 Subject: MATLAB issues Message-ID: <81563058df871773b1c3023bf47413a7.squirrel@webmail.andrew.cmu.edu> Dear Autonians, It has been brought to my attention that MATLAB licensing manager is down on several computing nodes lov3, lov4, low1, lot1. This is due to the fact that these machines were installed almost a year ago and license information have to be updated. As the release of MATLAB R2014b is immanent (I expect it any within next couple of days) I would rather wait to install new MATLAB R2014b than use my time to fix something which will be obsoleted by next Monday. In the mean time you can use LOT2 if you need MATLAB where the fresh copy was installed a two days ago. Most Kind Regards, Predrag From predragp at andrew.cmu.edu Fri Oct 3 00:28:34 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Fri, 3 Oct 2014 00:28:34 -0400 Subject: Observium Message-ID: <93018409769e619706d38c1a2955d9ca.squirrel@webmail.andrew.cmu.edu> Dear Autonians, I am announcing immediate availability of Observium, Network Observation and Monitoring System (NOMS) for Auton Lab. https://monit.autonlab.org/ username: auton password: Dr.Who Unlike M/Monit http://monit.autonlab.org username: auton password: Dr.Who which is good for quick up and down view of all our servers Observium strong side is providing long term network metric collection (into MySQL database) suitable for analytics. So far I have added only 5 devices as a technology preview but the fact that Observium uses SNMP for pooling will making adding remaining servers breeze (it takes about 15-20 minutes per server). I would still like to add the last element of Auton Lab monitoring triad, a centralized login server (probably syslog-ng). Observium has plugins for syslog-ng but it appears that integration is not smooth (waisted couple of hour trying to get collectd graphs appear in Observium but no avail). Most Kind Regards, Predrag Punosevac From predragp at andrew.cmu.edu Mon Oct 6 15:43:37 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Mon, 6 Oct 2014 15:43:37 -0400 Subject: MATLAB R2014b being installed Message-ID: <39431d12b59dc931451737ae56829064.squirrel@webmail.andrew.cmu.edu> Dear Autonians, As I anticipated MATLAB R2014b has been released this morning. The new version is already installed and tested on low1. I am coping installation files on all computing nodes at this time. I anticipate that by 9:00 PM all computing nodes and desktops will have working copy of MATLAB R2014b. The process is fully automatized and parallelized but it takes time due to the network load to move 8GB of files to 20 computing nodes and 16 desktops and install them. Most Kind Regards, Predrag From predragp at andrew.cmu.edu Tue Oct 14 08:56:00 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Tue, 14 Oct 2014 08:56:00 -0400 Subject: LOV3 crashed over night Message-ID: While you were sleeping somebody (lets not name the names) have crashed LOV3. It self rebooted. Obviously all work is lost. The same person have also filled the scratch directory :) Cheers, Predrag From predragp at andrew.cmu.edu Wed Oct 22 09:16:09 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Wed, 22 Oct 2014 09:16:09 -0400 Subject: Problems accessing CVS In-Reply-To: References: Message-ID: > Hi Predrag, > > I'm trying to check out some directories from CVS, but I'm getting the > following: > > "ssh: connect to host svnhub.int.autonlab.org port 22: No route to host" > > Any ideas why this is happening? I've checked out files without problems > in > the past. > Thanks, > Dylan Fitzpatrick > While I was sleeping ssh daemon on svnhub died for unknown reasons to me. Dylan was apparently the first one to see it. I hope that it is just that because I still see machine up and running on M/Monit. Nothing can be done until I get into the machine room and physically inspect the server. Please give me at least 2h before you hear from me again. Predrag From predragp at andrew.cmu.edu Wed Oct 22 10:56:24 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Wed, 22 Oct 2014 10:56:24 -0400 Subject: Problems accessing CVS FIXED! In-Reply-To: References: Message-ID: <9d009c325eb798a3c08d53ac69c74c96.squirrel@webmail.andrew.cmu.edu> >> Hi Predrag, >> Fixed! Predrag >> I'm trying to check out some directories from CVS, but I'm getting the >> following: >> >> "ssh: connect to host svnhub.int.autonlab.org port 22: No route to host" >> >> Any ideas why this is happening? I've checked out files without problems >> in >> the past. >> Thanks, >> Dylan Fitzpatrick >> > > While I was sleeping ssh daemon on svnhub died for unknown reasons to me. > Dylan was apparently the first one to see it. I hope that it is just that > because I still see machine up and running on M/Monit. Nothing can be done > until I get into the machine room and physically inspect the server. > Please give me at least 2h before you hear from me again. > > Predrag > > From predragp at andrew.cmu.edu Wed Oct 22 22:57:45 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Wed, 22 Oct 2014 22:57:45 -0400 Subject: ViewVC migrated to Apache Message-ID: Dear Autonians, Many of you like viewing the contents of CVS and SVN repositories using a web browser. As most of you probably know we use ViewVC an open source tool for viewing the contents of our CVS and SVN repositories. ViewVC was chosen 2009 after an unsuccessful attempt to migrate Auton code from CVS to Subversion server due to its unique feature to support both CVS and SVN version control systems. CVS was not retired due to reliance on ancient gmake-magic which doesn't support SVN (since SVN didn't exist 20 years ago). Many of you have notice that since the CVS and SVN repos have been migrated to svnhub.int.autonlab.org VeiwVC http://svnhub.int.autonlab.org/viewvc (from Auton network only or using proxy from outside the lab) was crashing a lot. This was due to the fact that built in python web server which comes with ViewVC has serious memory leak. Tonight I am pleased to report that ViewVC has been migrated to Apache and the you should not see any more crashes. An additional bonus of migration is that I enabled CvsGraph graphical CVS revision tree generator. Please check out and see CVS diffs as trees :) Most Kind Regards, Predrag P.S. On the long term use of both CVS and SVN as well as "magic" obsolete packages from /auton/software is suboptimal. A proper solution is migration from gmake-magic which works only with CVS to cmake for example. The consolidation of all repos under one version control system which will be well integrated with Web viewer but more importantly with bug tracker so that bug reports and be tagged to the code (right now our Mantis bug tracker is not integrated with our code repositories and very rarely used by half dozen people). From predragp at andrew.cmu.edu Fri Oct 24 12:34:16 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Fri, 24 Oct 2014 12:34:16 -0400 Subject: Life Preserver Message-ID: Dear Autonians, Most of you by now know that we are using the most advanced file system in existence ZFS on our file servers GAIA and Neill-ZFS. Some of you might have bothered to Google about ZFS and might have noticed that ZFS supports journaling of course via periodic snapshots. It is my pleasure to announce that this feature is now fully production ready and turned on. Periodic snapshots of your home directories are taken every day between hours 8:00 AM and 11:00 PM on every hour and kept for two weeks. Periodic snapshots of project, data, and neill-fs ZFS data sets are taken Monday through Friday between 9:00 AM and 6:00 PM and kept for four weeks. To put in layman's terms what I am telling you is that if you accidentally delete file/directory or want to reverse changes in a file which is located on a file server (/zfsauton or /zfsneill folders) don't despair. In all likelihood I have your file in ZFS journal (snapshot) and all you need to do is e-mail me. I will be happy to recover your file for you. Most Kind Regards, Predrag Punosevac P.S. It would be quite possible to me to provide Life Preserver for your personal laptops but the orders have to come from my bosses. From predragp at andrew.cmu.edu Tue Oct 28 00:14:58 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Tue, 28 Oct 2014 00:14:58 -0400 Subject: LXV1 down Message-ID: <101007a225ec6abac0e86d161a3ffdee.squirrel@webmail.andrew.cmu.edu> Dear Autonians, LXV1 has apparently crashed taking with herself down 4 KVM virtual hosts. There is nothing I can do until I go tomorrow morning to NREC and check the physical condition of the server. I apologize for the inconvenience. Most Kind Regards, Predrag From predragp at andrew.cmu.edu Tue Oct 28 13:06:49 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Tue, 28 Oct 2014 13:06:49 -0400 Subject: LXV1 is clinically dead Message-ID: <13695d7d994b3ba71421746f5f720050.squirrel@webmail.andrew.cmu.edu> Dear Autonians, LXV1 appears to be clinically dead. I could not power up the machine. My hunch is that the power supply is burned. Until we receive new power supply from Silicon Mechanics LXV1 remains out of commission. NREC uses still have at their disposal LXV2 computing node as well as file server Lera. I have a clone of vibtrend virtual machine on LXV2 which can be fired up if necessary. Win-REC, and vibtrend-test virtual machines will remain down with LXV1. I apologize for the inconvenience. Most Kind Regards, Predrag From predragp at andrew.cmu.edu Tue Oct 28 23:37:41 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Tue, 28 Oct 2014 23:37:41 -0400 Subject: RHEL 6.6 kernel 2.6.32-504 Message-ID: <411137f94559975c0c87d0b202260ed6.squirrel@webmail.andrew.cmu.edu> Dear Autonians, 6th update of Red Hat 6 has been released on 13 October 2014 (kernel 2.6.32-504). Josko, Prentice and bunch of other slackers from Princeton and Rutgers needed whole two weeks to update Springdale Linux clone which we use in our lab to version 6.6. Consequently you have been deprived until now from this latest and greatest code. Your user-land is up to date now but in order to pick kernel 2.6.32-504 you will have to reboot the machine. Since computing nodes seem to work fine I personally have no plans to to force reboot them just to pick a minor kernel update. However I plan on rebooting my desktop. Some unimportant virtual hosts are also rebooted. You will also notice that Java is updated to R7u71 on all desktops, nodes and most virtual machines. I update VirtualBox to version 4.3-4.3.18_96516_el6-1.x86_64 on your desktops except in the cases VirtulBox was running. If you work with Apache-Tomcat make sure you pick the latest update 7.0.56. Public, vibtrend and few other instances of TCWI I am taking care of are updated. Finally if you work with MySQL it has been just updated to 5.6.21. I have only update tj-ondeck and tj-dev virtual machines. Sincerely, Predrag From predragp at andrew.cmu.edu Thu Oct 30 20:08:03 2014 From: predragp at andrew.cmu.edu (predragp at andrew.cmu.edu) Date: Thu, 30 Oct 2014 20:08:03 -0400 Subject: the navy server is fixed In-Reply-To: <5452A24E.8020100@cs.cmu.edu> References: <5452A24E.8020100@cs.cmu.edu> Message-ID: Yes! LXV1 was fixed at 4:30 PM. I have not had chance to send an e-mail earlier. Sorry I don't have a smart phone. Predrag From awd at cs.cmu.edu Fri Oct 31 17:43:17 2014 From: awd at cs.cmu.edu (Artur Dubrawski) Date: Fri, 31 Oct 2014 17:43:17 -0400 Subject: Auton Lab seminar on Wednesday November 5th, NSH 3305 at 1:30pm, with Kai Vetter Message-ID: <54540275.3040608@cs.cmu.edu> Title: New Concepts in Nuclear Radiation Detection Relevant for Research and Security -- Challenges and Opportunities for Autonomous Systems and Multi-Sensor Fusion Speaker: Kai Vetter Department of Nuclear Engineering, University of California, Berkeley Applied Nuclear Physics, Lawrence Berkeley National Laboratory Abstract: Recent developments in the detector fabrication, signal readout, and data processing enable new concepts in radiation detection that are relevant for applications ranging from fundamental physics to medicine and nuclear security. I will review some examples of recent efforts in our Berkeley Applied Nuclear Physics program on these technical developments and how they are mapped to questions in basic sciences, biomedical imaging, and nuclear security. The presentation will focus on the opportunities and challenges in fusing nuclear radiation with contextual and environmental data. Examples for applications include contamination assessment and mapping in Fukushima in Japan, emergency response and consequence management, detection, as well monitoring, for example of nuclear processes inside facilities. Not only need the sensitivity for nuclear signatures be increased, they need to be correlated with contextual and environmental information to enable the sensitivity and information needed to act or to plan and guide operations, sometimes in realtime. Host: Artur Dubrawski (awd at cs.cmu.edu) Appointments: Karen Widmaier (krw at andrew.cmu.edu) Place and time: NSH 3305, 1:30pm Wednesday, November 5th. -------------- next part -------------- An HTML attachment was scrubbed... URL: