Discuss

discuss@lists.centos.org

34753 discussions

x86_64 EDAC throwing error
by Kurian Thayil 17 Jul '09

17 Jul '09

Hi All, We have installed CentOS 5.3 x86_64 in an HP DL585 server with AMD Opteron 64 bit processor and 16 GB RAM. The kernel version is 2.6.18-128.el5 . Now this has thrown an error message in /var/log/message, Jul 3 21:41:11 db1 kernel: EDAC k8 MC0: general bus error: participating processor(local node origin), time-out(no timeout) memory transaction type(generic read), mem or i/o(mem access), cache level(generic) Jul 3 21:41:11 db1 kernel: EDAC MC0: CE page 0x65bc7, offset 0x6a0, grain 8, syndrome 0x6e1a, row 0, channel 0, label "": k8_edac Jul 3 21:41:11 db1 kernel: EDAC k8 MC0: extended error code: ECC chipkill x4 error Jul 3 22:00:00 db1 ntpdate[3813]: step time server 120.88.46.10 offset -4.375417 sec Jul 3 22:12:57 db1 kernel: EDAC k8 MC0: general bus error: participating processor(local node origin), time-out(no timeout) memory transaction type(generic read), mem or i/o(mem access), cache level(generic) Jul 3 22:12:57 db1 kernel: EDAC MC0: CE page 0x65bc7, offset 0x6a0, grain 8, syndrome 0x6e1a, row 0, channel 0, label "": k8_edac Jul 3 22:12:57 db1 kernel: EDAC k8 MC0: extended error code: ECC chipkill x4 error I understand this is an error message from Error Detection And Control module and just wanna confirm that this is not a kernel or software related issue. If Hardware related is it confined only to the Physical memory stick installed or processor related? Any hint on this?? Regards, Kurian Thayil.

4 3

NIC traffic monitoring, recording and reporting software?
by James B. Byrne 16 Jul '09

16 Jul '09

I have a remote CentOS-5.3 installation that shares traffic with another enterprise. I want to record and measure the actual network traffic that comes into and goes out of that host's net i/f on a monthly basis for billing verification purposes. Does CentOS have a software package to accomplish this? I have run across references to a package called vnstat as a possible answer to my requirements, but I cannot seem to locate a CentOS rpm package for it. Regards, -- *** E-Mail is NOT a SECURE channel *** James B. Byrne mailto:ByrneJB@Harte-Lyne.ca Harte & Lyne Limited http://www.harte-lyne.ca 9 Brockley Drive vox: +1 905 561 1241 Hamilton, Ontario fax: +1 905 561 0757 Canada L8E 3C3

6 10

CentOS-announce Digest, Vol 53, Issue 5
by centos-announce-request＠centos.org 16 Jul '09

16 Jul '09

Send CentOS-announce mailing list submissions to centos-announce(a)centos.org To subscribe or unsubscribe via the World Wide Web, visit http://lists.centos.org/mailman/listinfo/centos-announce or, via email, send a message with subject or body 'help' to centos-announce-request(a)centos.org You can reach the person managing the list at centos-announce-owner(a)centos.org When replying, please edit your Subject line so it is more specific than "Re: Contents of CentOS-announce digest..." Today's Topics: 1. CESA-2009:1154 Critical CentOS 3 i386 dhcp - security update (Tru Huynh) 2. CESA-2009:1154 Critical CentOS 3 x86_64 dhcp - security update (Tru Huynh) ---------------------------------------------------------------------- Message: 1 Date: Wed, 15 Jul 2009 21:59:01 +0200 From: Tru Huynh <tru(a)centos.org> Subject: [CentOS-announce] CESA-2009:1154 Critical CentOS 3 i386 dhcp - security update To: centos-announce(a)centos.org Message-ID: <20090715195901.GA20062(a)sillage.bis.pasteur.fr> Content-Type: text/plain; charset="us-ascii" CentOS Errata and Security Advisory CESA-2009:1154 dhcp security update for CentOS 3 i386: https://rhn.redhat.com/errata/RHSA-2009-1154.html The following updated file has been uploaded and is currently syncing to the mirrors: i386: updates/i386/RPMS/dhclient-3.0.1-10.2_EL3.i386.rpm updates/i386/RPMS/dhcp-3.0.1-10.2_EL3.i386.rpm updates/i386/RPMS/dhcp-devel-3.0.1-10.2_EL3.i386.rpm source: updates/SRPMS/dhcp-3.0.1-10.2_EL3.src.rpm You may update your CentOS-3 i386 installations by running the command: yum update dhcp Tru -- Tru Huynh (mirrors, CentOS-3 i386/x86_64 Package Maintenance) http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xBEFA581B

1 0

nfs (v3?) fails to allow writes and permission changes
by Rob Kampen 16 Jul '09

16 Jul '09

Hi All, I have an nfs mount from my CentOS 5.3 client to a CentOS5.3 server. This has been functioning correctly for some years. Beginning last week I had problems saving OpenOffice documents onto the nfs mounted volume. OpenOffice just times out with a file io error. So I tried just a simple shell - cd to the mounted volume did a vi create of text file - no problems. tried a chmod +w on an owned file - returned okay but no change to the permissions when checked with ls -l tried sudo chmod +w on the same file and get [rkampen@robsws p_494]$ sudo chmod +w 5887_cover.pdf Password: chmod: changing permissions of `5887_cover.pdf': Operation not permitted yet using gnome file browser I can change permissions on these nfs mounted files just fine - go figure. This did work in the past - so what has changed? how do I check what version of nfs is actually working? [rkampen@robsws p_494]$ mount 192.168.230.230:/NDG on /NDG type nfs (rw,noatime,intr,addr=192.168.230.230) Any thoughts? Thanks Rob

6 18

Initial backup prior to daily rsync backups
by Jeff Gregor 16 Jul '09

16 Jul '09

I have a few CentOS 5.3 servers out at remote sites that connect over a satellite link (a painfully slow satellite link). There's a time period in the wee hours of the morning when the satellite provider doesn't stiff me for bandwidth, and when the link is generally idle, where I want to use rsync to back up files to my local server. I've got rsync working in in combination with cron and ssh, and that process works fine. The problem is getting the initial copy of the files over to the local server so that rsync can just deal with files changed during the day (there tend to not be too many changes -- mostly office-type documents, few pictures, that sort of thing, so limited bandwidth is not a problem). If I run rsync now, however, it is trying to copy over all the files since none exist on the local servers, and by my calculations copying that 160gb worth of data should be done some time in mid-February, 2010. I can make the 4-hour round trip out to the remote site with a USB hard drive and copy the files that way, but my concern is how to get the file structure, timestamps, permissions, etc copied exactly from the remote server > USB drive > local server so that rsync can run most efficiently (and not have to do a compare on every single file the first time it runs.) So the question is, how do I copy the files from the remote server back to my local server in a way that will allow rsync, once the copy is complete, to start dealing just with the daily changes in the most efficient way possible? Thanks!

3 3

OT: Web "Project" type software
by James B. Byrne 16 Jul '09

16 Jul '09

On Wed Jul 15 18:37:03 UTC 2009, James Pifer jep at obrien-pifer.com wrote: > Sorry for the OT... I'm looking for some software to fill a > fairly specific set of requirements. I'm not necessarily looking > for project software, but it seems like the closest to match what > I need. I'm also not tied into open source or free. If anyone > knows of any software that might fill these requirements I would > appreciate it. I am building a fairly sizable dbms based project and I started out with subversion as the SCM and Trac as the PAS. Last year I switched the project to Git and Redmine respectively. Both are open source. Git is the Linux kernel maintainers' SCM system and was initially designed and written by LT himself to replace BitKeeper. It is a completely distributed SCM and is very easy to use, albeit not so easy to learn or understand. > Here's what I'm looking for: > 1) Web enabled application Redmine is 100% Ruby and Rails, totally web based. It supports Git repositories out of the box. Git is a combination of scripts and C programs. It is available for Linux, Microsoft and several other OS platforms. Repository structure is identical for all platforms although differs in implementation details. On MicroSoft environments I hold that Git is best hosted under a cygwin instance, although a native port does exist. > 2) File storage - need to stare a large amount of large cad files - be able to categorize file storage (not one large list) - keep different revisions of files with some kind of notes - be able to search files (at least include names in a search) Remine uses the underlying fs for file storage and provides the metadata for search. Files can be placed into a document list or attached to individual wiki pages. A description can be provided upon creation, although at present there seems no supported way of editing an existing description. Redmine supports full test searching. However, Git is by far and away the best solution to the problem of large file storage. Redmine can index a Git repository and track individual change sets. Therefore, Git + Redmine will give you most, if not all, of what you ask for. > 3) Task - assign and keep track of tasks (searchable) Redmine supports multiple projects and sub-projects in a single instance. Issues may be related to one another. Users may be assigned differing roles on different projects. Different projects can have different categories and statuses. The DBMS backend is PostgrSQl so any report writer with an SQL backend should be able to provide whatever reports that you have the data to drive. > 4) Store meeting notes and minutes (searchable) In addition to issue tracking and file storage, Redmine provides both project specific wikis and web forums. Access to either can be controlled on a project/user basis. The full text search facility covers both. It also has time logging and a slew of other features. > 5) Would like to support LDAP I believe that Redmine does, but I do not use this myself so I cannot be sure. Git is a filesystem, it has its own network transfer protocol which is very fast and completely insecure. It also supports transfer over http, very slow and insecure, and ssh, acceptable speed and secure transfer. > 6) Extension on file storage - since we'll be storing a large files > and many of our users are across a WAN we'd like to sync/mirror > the files to a few different servers. I'm not sure how we'd control > how the users get the "closest" server but.... If you use Git instead of Redmine for your large file store then each user gets a complete repository the first time that they pull (clone actually) from the remote, including all the source changes to initialisation. Thereafter they only get the diff objects. Git uses a very efficient storage mech. that basically only passes compressed diffs of file contents. It uses MD5 hashes of file contents to identify change sets so that if you rename a file or merge two into one it can track these events and avoid duplicate storage. I use a cron scheduled git pull over ssh to maintain a remote copy of our base repository and the traffic created is negligible even after large commits. If the Git repository containing those files is bound to your Redmine project then you can search the files therein through Redmine and compare diffs across multiple versions. Of course, nothing is going to help much with binary data files. HTH -- *** E-Mail is NOT a SECURE channel *** James B. Byrne mailto:ByrneJB@Harte-Lyne.ca Harte & Lyne Limited http://www.harte-lyne.ca 9 Brockley Drive vox: +1 905 561 1241 Hamilton, Ontario fax: +1 905 561 0757 Canada L8E 3C3

1 0

Re: [CentOS] My server reboots every hour! Help please! [update]
by Marko Vojinovic 16 Jul '09

16 Jul '09

On Wed, Jul 15, 2009 at 1:16 PM, Marko Vojinovic<vvmarko(a)gmail.com> wrote: > I have a machine here that resets itself every one hour (without my > intention, of course): Ok, there has been some development of the situation. I asked a collegue of mine (who happens to have physical access to the machine) to shut it down and get into the BIOS to check for the "wake on lan" status. He did (it was off) and he left the machine hanging in BIOS for an hour waiting to see if it would restart. It didn't, for two hours. He was also monitoring temerature and voltage and comparing them with two other machines in the room (identical hardware) and nothing seemed out of ordinary. Then he rebooted to CentOS, and since then restarts have not happened for a whole day now: # uptime 16:26:25 up 22:11, 1 user, load average: 0.00, 0.00, 0.00 That means that (a) the problem went away and (b) I am not able to reproduce it. From the logs I read that the resets were happening every 3625 +- 10 seconds, so to speak. The idea of intentionally skewing the system clock and wait for a reaction was my next step, but since it stopped restarting I cannot verify anything now. Since the problem went away, I am moderately happy, so we can drop the thread, and I'll make sure to reopen it if I see it restarting again. Thanks to all for suggestions and help! Best, :-) Marko

1 0

Keep lossing wireless - ipw3945
by Robert Moskowitz 16 Jul '09

16 Jul '09

Sunday I updated my kernel to 128.1.16 (I had to reboot so it was time to update). No problems with wireless in my home. Monday I started traveling to IEEE 802 plenary, and I have lost my wireless numerous times needing to reboot. The symptom is the radio light goes off. Back before the kernel update, this would happen sometimes at home while on battery, and I would just press the capacitor sensor and turn it back on, and I would get my signal back. But so far this week, this has not worked. Nothing. No signal is gotten after turning the radio back on (for whatever reason this system decides to turn it off). I have tried restarting the network and NetworkManager and this has not helped (this use to help in the past when I saw this behaviour). There were some messages during the kernel update as follows: Uninstalling: fuse 2.7.4-1.nodist.rf (2.6.18-128.1.6.el5) (i686) -------- Uninstall Beginning -------- Module: fuse Version: 2.7.4-1.nodist.rf Kernel: 2.6.18-128.1.6.el5 (i686) ------------------------------------- Status: Before uninstall, this module version was ACTIVE on this kernel. Removing any linked weak-modules WARNING: Can't read module /lib/modules/2.6.18-128.1.6.el5/weak-updates/fuse.ko: No such file or directory WARNING: Can't read module /lib/modules/2.6.18-128.1.6.el5/weak-updates/fuse.ko: No such file or directory fuse.ko: - Uninstallation - Deleting from: /lib/modules/2.6.18-128.1.6.el5/extra/ - Original module - No original module was found for this module on this kernel. - Use the dkms install command to reinstall any previous module version. depmod.... DKMS: uninstall Completed. Uninstalling: ipw3945 1.2.1-1.nodist.rf (2.6.18-128.1.6.el5) (i686) Error! The module ipw3945 1.2.1-1.nodist.rf is not currently installed. This module is not currently ACTIVE for kernel 2.6.18-128.1.6.el5 (i686). Cleanup : kernel [6/6] Removed: kernel.i686 0:2.6.18-128.1.6.el5 kernel-devel.i686 0:2.6.18-128.1.6.el5 Installed: kernel.i686 0:2.6.18-128.1.16.el5 kernel-devel.i686 0:2.6.18-128.1.16.el5 Updated: kernel-headers.i386 0:2.6.18-128.1.16.el5 Complete! But despite that unistalling ipw3945 message, wireless worked, or it seemed. I am seeing log messages like: Jul 15 10:35:28 nc2400 kernel: ipw3945: Detected geography ABG (11 802.11bg chan nels, 13 802.11a channels) a lot. When I lost the wireless and tried to restart everything I saw: Jul 15 10:37:27 nc2400 NetworkManager: <WARN> scan_results_cb(): could not get scan results: An error ocurred getting scan results.. I really cannot deal with frequent rebooting. Can someone point me in a direction to fix this?

4 7

Re: [CentOS] Conflicting perl packages?
by Les Bell 16 Jul '09

16 Jul '09

Marko Vojinovic <vvmarko(a)gmail.com> wrote: >> The simple ordinary yum update of CentOS 5.3 spits a bunch of transaction check errors regarding packages perl-IO-Compress-Zlib-2.015-1.el5.rf.noarch and perl-IO-Compress-2.020-1.el5.rf.noarch which is supposed to replace perl-IO-Compress-Base-2.015-1.el5.rf.noarch << I had the same problem on one system, which was niggling at the back of my mind. I fixed it by brute force: rpm -e --nodeps perl-Compress-Zlib perl-IO-Compress-Zlib yum install perl-IO-Compress Ugly, but it works. I also noticed some messages when restarting amavisd-new, relating to error loading optional module . . . DKIM.pm, which *might* be related (hadn't noticed it before, but it was also appearing on another system which didn't have the Zlib errors above). Easily fixed with yum install perl-Mail-DKIM I did this half an hour or so ago, and everything seems to be running smoothly. Filipe Brandenburger <filbranden(a)gmail.com> wrote: >> yum --skip-broken update << I tried that, as well as setting skip_broken in /etc/yun.conf, but the Zlib errors aborted the entire transaction every time, forcing me to update other packages one by one. That's what motivated me to do the hack above. Best, --- Les Bell [http://www.lesbell.com.au] Tel: +61 2 9451 1144

1 0

CentOS/SNMP update breaks MRTG?
by Noob Centos Admin 16 Jul '09

16 Jul '09

I got itchy fingers over the weekend and decided to fix what wasn't broken and upgraded one of the older servers from Centos 5.2 to Centos 5.3. Following the recommended process of updating glibc and such before the rest, it appeared to work perfectly and rebooted without problem. However, MRTG 2.15.2 started complaining about unexpected values. I installed/updated both MRTG (2.16.2) and net-snmp to the latest available in hope of fixing it. Subsequently, MRTG stopped working altogether. I've spent the whole weekend and whole Monday morning trying to fix it and thus far have only finally managed to get garbage values showing up in MRTG again as opposed to nothing. And this required learning about SNMP and adding many additional lines to the original MRTG configuration file, none of which I had to do previously. Did anybody else have similar experiences with MRTG failing after the update and what was the simple fix? It does not make any sense that I have to jump through so much hoops to get just the default functionality back. Thus I believe there must be one small thing I'm overlooking. Thanks for any advice.

7 16

Jump to page:

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Discuss