Discuss

discuss@lists.centos.org

34753 discussions

kABI-tracking kmod-xfs for CentOS-4 (Was: Preventing hour-long fsck on ext3-filesystem)
by Akemi Yagi 18 Jun '09

18 Jun '09

On Tue, May 26, 2009 at 1:51 PM, Tru Huynh <tru(a)centos.org> wrote: > On Tue, May 26, 2009 at 08:16:42AM -0700, Akemi Yagi wrote: >> Here is the updated version: >> >> http://centos.toracat.org/kmods/CentOS-4/xfs/SRPMS/ >> >> Please discard the obsoleted ones (I did not bump the version/release >> number). Let me know when your binaries are ready for testing. >> > rebuilds and uploaded. * Downloaded from http://people.centos.org/tru/kABI/4/RPMS/ and did some test installs. First, installed the kernel version specific kmod-xfs-0.4-1.el4.2.6.9_78.0.22 to the test box and then installed the current kABI-tracking kmod-xfs-0.4-2.el4. It successfully replaced the xfs.ko installed by the former and created symlinks to other installed kernels correctly. Second, installed the current kmod-xfs to an older kernel and then updated the kernel. Upon reboot, a new symlink was created for the newly installed kernel successfully. Third, confirmed that the kABI-tracking kmod-xfs recognized centosplus kernels as expected (even -vm kernels). One example of such symlinks: $ uname -mr 2.6.9-78.0.22.plus.c4 i686 $ ls -l `find /lib/modules -name xfs.ko` lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-67.0.1.EL.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-67.0.22.ELvm/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-67.0.4.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.0.1.ELvm/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.0.1.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko -rw-r--r-- 1 root root 598576 May 26 13:07 /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:34 /lib/modules/2.6.9-78.0.22.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.0.5.EL/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.0.5.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.0.8.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko lrwxrwxrwx 1 root root 46 May 26 18:20 /lib/modules/2.6.9-78.plus.c4/weak-updates/xfs/xfs.ko -> /lib/modules/2.6.9-78.0.22.EL/extra/xfs/xfs.ko * I can only test installs/uninstalls. If someone can test it by actually *using* it, that will be great. I did not touch the code, so it should work. Nevertheless... * Some side note In the current kmod-xfs, modules are partially strip'ed. This has a dramatic effect on the size of the modules (hence the packages themselves). For example, the obsoleted rpm's were this big: 3413556 May 25 16:50 kmod-xfs-0.4-2.el4.i686.rpm 3500186 May 25 16:50 kmod-xfs-0.4-2.el4.x86_64.rpm The current packages are: 299042 May 26 13:23 kmod-xfs-0.4-2.el4.i686.rpm 308712 May 26 13:23 kmod-xfs-0.4-2.el4.x86_64.rpm All in-kernel modules are strip'ed during compilation. While it is not required, I think it's better to process the kmods the same way. Akemi

3 4

CIFS Issue When Copying Large/Many Files From CentOS To Remote Windows 2003 Server Share
by Kemp, Larry 18 Jun '09

18 Jun '09

My first post to the list, hope the community can help me, I am having a problem with CIFS:VFS. I am trying to write large files (and a lot of files) from my Cent OS 5.3 server to a remote Windows 2003 Server share. It appears there might be some incompatibilities between how the two OS's individually implement CIFS, maybe. I automount the remote share via /etc/fstab and have the /mnt/remotewinservershare built. Everything is setup and seems to work properly but during the large file transfer (at intermittent times) the Linux Kernel seems to panic and the CIFS connection seems to fail as a result since CIFS happens at the Kernel level, returning errors. Although when I cd the /mnt/remotewinservershare dir, or ls /mnt the mount is there is I can manually cp to it with absolutely no problems. My /etc/fstab line is: //10.100.101.205/share /mnt/remotewinservershare cifs username=MYUSER,pass=MYPASSWD 0 0 The CIFS errors returned are: WARNING: Kernel Errors Present CIFS VFS: Error -4 sending data ...: 103 Time(s) CIFS VFS: Unexpected lookup error -112 ...: 2 Time(s) CIFS VFS: Error -4 sending data ...: 380 Time(s) I have seen in various online forums people have run into this since 2004 and posts in 2007 have gone unanswered. Seems to be isolated to Cent OS trying to CIFS to a Windows 2003 Server when transferring large or many files (and does so intermittently). Has anyone has hit this before and been able to fix this please share how you worked around it or fixed it. It is important for me to be able to have a stable CIFS connection to this remote Windows Server 2003 server, and we know we are not getting a look at the Windows kernel. Big Thanks LK

2 3

Another Postfix (or procmail ??) question
by Michael A. Peters 18 Jun '09

18 Jun '09

Is there a simple way to strip out what I assume is a header that asks for a receipt that the message was delivered? I never like to send such receipts, and they are annoying. It mostly happens on a couple lists I'm on (not this one) - I don't just want it for myself, I may be setting up a list soon and if soon, would want to strip such requests out of the message before the list server sends it to the list.

2 1

script help
by chloe K 18 Jun '09

18 Jun '09

Hi I have a file. list.txt (two columns) column1 column2 name address I need to put in the letter file letter.txt eg: Dear: Chloe Address: CA Can I use this for i `cat list.txt` | sed 's/Chloe/$i.1; /CA/$i.2/g' $i.letter.txt Thank you for your help __________________________________________________________________ Looking for the perfect gift? Give the gift of Flickr! http://www.flickr.com/gift/

9 10

Howto change arch for FirebirdSS rpm
by Pavel Lisy 18 Jun '09

18 Jun '09

Hello I have this problem: I make my own php rpms, base on src.rpm from CentOS but with my patches (more options included). I have built it on my machine originally but now I want make it in mock environment. I solved almost everything but Firebird support for 64bit arch. Original RPM from http://www.firebirdsql.org/ are made with amd64 arch but my php is x86_64. And I don't know how to solve it. .src.rpm for FirebirdSS and/or FirebirdCS is on available. I have this section in .spec %if %{with_interbase} %package interbase Group: Development/Languages Prereq: php = %{version}-%{release}, perl Summary: A Sybase/MSSQL database module for PHP. Provides: php_database # BuildPrereq: /usr/include/gds.h /usr/include/ibase.h # BuildPrereq: /usr/include/iberror.h /usr/include/ib_util.h # BuildPrereq: FirebirdCS >= 1.5.0 # "AutoReqProv: yes" don't work because of fault in Firebird RPMs # # AutoReqProv: no # BuildPrereq: FirebirdSS >= 1.5.0 # # Requires: /usr/lib/libgds.so %ifarch i386 i486 i586 i686 BuildPrereq: FirebirdSS >= 1.5.0 %endif %ifarch x86_64 BuildPrereq: FirebirdSS.amd64 >= 2.0.1 %endif But mock failed with this: DEBUG util.py:256: No Package Found for FirebirdSS.amd64 >= 2.0.1 │54 I don't how to include amd64 arch rpm to my x86_64 repository. (my repo: http://ftp-hk.tmapy.cz/tmapy-twist/centos/4/os/x86_64/RPMS/) I think there are two possible solution, rebuilt Firebird rpm with different arch or include amd64 to x86_64 repository. Do you how to make it? My questions: 1. Can I reconstruct .src.rpm from binary rpm or change arc in binary arch? How? 2. How to include amd64 arch rpm to my x86_64 repository? Thanks in advance Pavel

1 0

Rsync, SSH and authorized_keys problem
by Stewart Williams 18 Jun '09

18 Jun '09

Hi, I'm trying to backup from one machine to the other (automatically via cron) using rsync and ssh password-less public key authentication. I having been trying to set this up following an article in a Linux magazine[1] by only allowing the specific rsync command to run on the remote box. I am using the following rsync command: $ rsync -avz -e "ssh -i ~/.ssh/backup-key" /backup stewart@name.of.remote.server:/backup This runs, connects using keys asking for no password and completes successfully until I add the above command to my authorized_keys file on the remote box: command="rsync -avz -e "ssh -i ~/.ssh/backup-key" /backup stewart@name.of.remote.server:/backup" ssh-dss ... key ... The client then says it cannot find the key, so the connection fails. Is it because it's trying to find the private key in the ~/.ssh directory on the remote box? Is the article wrong? Or am I doing something wrong? Should I use the $SSH_ORIGINAL_COMMAND variable? Regards, Stewart Williams [1] http://www.linuxformat.com/pdfs/download.php?PDF=LXF105.tut_backup.pdf

7 11

Cron Mail
by Joseph L. Casale 18 Jun '09

18 Jun '09

Is there any way to prevent mail from completed cron jobs for only say the hourly directory? Thanks! jlc

2 2

Cyrus IMAP and ipurge
by Spiro Harvey 18 Jun '09

18 Jun '09

Anyone here familiar with Cyrus IMAP and its util ipurge? We'd like to set auto deletion of our users' spambox folders, but I'm a bit paranoid about its functionality. The main things that confuse me are from the man page: "Ipurge deletes messages from the mailbox(es) specified by mailbox-pattern" with no indication as to what defines "mailbox-pattern" Second: "Ipurge by default only deletes mail below shared folders, which means that mails in mail- box(es) below INBOX.* and user.* stay untouched. Use the option -f to also delete mail in mailbox(es) below these folders." What's a "shared folder"? OK, so the format of a Junk mail folder is: user/spiro/Junk(a)knossos.net.nz and I've successfully done: $ ipurge -f -d 90 user/spiro/Junk(a)knossos.net.nz Working on user/spiro/Junk(a)knossos.net.nz... total messages 21 total bytes 187742 Deleted messages 16 Deleted bytes 138324 Remaining messages 5 Remaining bytes 49418 however, without the -f, it did nothing, even though I was explicitly supplying a folder. Clearly I misunderstand what the man page is trying to tell me. I'd like to set up a proper "pattern" to purge all spamboxes correctly via the cyrus.conf, because at the moment, I've got an awk script that generates a list of email addresses and produces a list in the format of user/userid/Junk@domain. I could then go over that list and run an ipurge for each one, but if anything happens that causes that list to be corrupted, I'm afraid of deleting the wrong mail. or worse. :( according to what I've gathered, the line: purgespam cmd="ipurge -f -d 90 user/%/Junk@*" at=0430 should work when put in cyrus.conf's Events section, but I can't confirm these wildcards. Some results on google just refer to * for the mailbox name, but none seem to include the domain as part of the string. Can anyone here shed some light? -- Spiro Harvey Knossos Networks Ltd 021-295-1923 www.knossos.net.nz

5 9

x86_64 CentOS 5.3 Users - Library Issue
by Justin Bull 18 Jun '09

18 Jun '09

Hello, I need to conduct a little research. Anyone who is running CentOS 5.3 x86_64 please run the following: # yum install libxml2-devel # nm /usr/lib64/libxml2.a | grep xmlXPathContextSetCache If my suspicions are correct, if you run cPanel/WHM you will not see any output and if without cPanel/WHM you'll see something similar to this: 00000000000023a0 T xmlXPathContextSetCache Thanks for your time and help. -- Best Regards, Justin Bull http://www.sohipitmhz.com/pubkey.txt (Public Key)

3 3

CentOS security advisories
by Joshua Bahnsen 18 Jun '09

18 Jun '09

I have been looking at the security advisories provided here: http://lists.centos.org/pipermail/centos-announce/ It appears that there is not a 1:1 correlation between advisories listed here and advisories listed by Red Hat: https://rhn.redhat.com/errata Is there a specific reason for this? Also, is there an alternate location to find all Errata information for CentOS? Joshua Bahnsen

4 13

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