Bug#575537: readahead-fedora: doesn't create later.sorted
Ok, so readahead-fedora should conflicts: auditd, but not if upstart is installed (but the conflicts mechanism can't handle that). Hmm. Can it fire off a log message or something if readahead detects that auditd may be clobbering it? Maybe by alerting if it doesn't find the usual /etc/init.d/stop-readahead-fedora in the data. (Which I guess should be a fairly clear indication that *something* went wrong.) Mike Stone -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
retitle 575537 readahead-fedora: collector conflicts with auditd thanks On Monday 03 May 2010 21:05:53 Michael Stone wrote: After looking at the debug info it seems that the logging stops right after auditd is started. So the early/later stuff was a red herring, and everything works fine if I remove auditd. So probably either readahead should conflict with that package, or something should change to make them work together. (I guess that it makes sense that they might have issues, since they're both trying to use the audit subsystem.) Yes, auditd conflicts with readahead's collector because it removes all the audit rules at startup. One way to workaround the conflict is by starting auditd after stop-readahead- fedora (which can not be easily done via the LSB init headers because of the dependency on $all, so the stop- script would have to call /etc/init.d/auditd start.) But that has the side effect that some audit events are not logged (actually, at present time auditd is incorrectly being started at rc2 instead of rcS, but that's a bug in the script's LSB header.) The alternative is to let the collector use the debugfs like ureadahead does, but requires a kernel patch. It's on my TODO list already but I don't know when I'm going to work on that. P.S. the conflict with auditd is documented in /usr/share/doc/readahead- fedora/README. Cheers, -- Raphael Geissert - Debian Developer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
After looking at the debug info it seems that the logging stops right after auditd is started. So the early/later stuff was a red herring, and everything works fine if I remove auditd. So probably either readahead should conflict with that package, or something should change to make them work together. (I guess that it makes sense that they might have issues, since they're both trying to use the audit subsystem.) Mike Stone -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
Sorry for the late response. I went on VAC and forgot about it. On 27 March 2010 09:38, Michael Stone mst...@debian.org wrote: On Fri, Mar 26, 2010 at 09:52:29PM -0600, Raphael Geissert wrote: * Non-comment contents of /lib/init/rw/readahead MOUNTED_USR=false MOUNTED_VAR=false Looks fine. * ls -l /etc/readahead.d -rw-r--r-- 1 root root 8760 2010-03-26 12:45 custom.early -rw-r--r-- 1 root root 10266 2010-03-27 07:43 early.sorted (there was a later.sorted yesterday, but I deleted it at some point to see if that would make a difference. Ran profile boot a half dozen times, deleted the file about halfway through. At the point I deleted it, it was 0 bytes with an mtime of Oct 2009 or so.) Strange. Could you please attach /etc/readahead.conf? RAC_LISTSEP should be set to /etc/init.d/later-readahead. * ls -l /etc/rc*.d/*readahead* lrwxrwxrwx 1 root root 25 2009-10-25 15:26 /etc/rcS.d/S02early-readahead - ../init.d/early-readahead* lrwxrwxrwx 1 root root 25 2010-03-26 10:03 /etc/rcS.d/S26later-readahead - ../init.d/later-readahead* Looks fine. Could you please turn on the collector's debug mode? (RAC_DEBUG=on in /etc/readahead.conf) Reboot afterwards in profile mode and attach (or upload somewhere, depending on how you are concerned about the privacy of the data) /dev/.readahead-rac.debug Thanks. Cheers, -- Raphael Geissert - Debian Developer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
On Fri, Mar 26, 2010 at 09:52:29PM -0600, Raphael Geissert wrote: * Non-comment contents of /lib/init/rw/readahead MOUNTED_USR=false MOUNTED_VAR=false * ls -l /etc/readahead.d -rw-r--r-- 1 root root 8760 2010-03-26 12:45 custom.early -rw-r--r-- 1 root root 10266 2010-03-27 07:43 early.sorted (there was a later.sorted yesterday, but I deleted it at some point to see if that would make a difference. Ran profile boot a half dozen times, deleted the file about halfway through. At the point I deleted it, it was 0 bytes with an mtime of Oct 2009 or so.) * sh -x /etc/cron.daily/readahead + TYPES='early later' + READAHEAD_CMD=/sbin/readahead + READAHEAD_BASE=/etc/readahead.d + MERGE_LISTS=false + '[' -x /sbin/readahead ']' + '[' -z '' ']' + export RENICED=yes + RENICED=yes + exec /sbin/start-stop-daemon --start -N +19 -I idle --exec /sbin/readahead --startas /etc/cron.daily/readahead * ls -l /etc/readahead.d (after running the cron script by hand) -rw-r--r-- 1 root root 8760 2010-03-26 12:45 custom.early -rw-r--r-- 1 root root 10266 2010-03-27 09:33 early.sorted * ls -l /etc/rc*.d/*readahead* lrwxrwxrwx 1 root root 31 2010-03-26 12:08 /etc/rc0.d/K01stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 12:08 /etc/rc1.d/K01stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 10:03 /etc/rc2.d/S23stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 10:03 /etc/rc3.d/S23stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 10:03 /etc/rc4.d/S23stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 10:03 /etc/rc5.d/S23stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 31 2010-03-26 12:08 /etc/rc6.d/K01stop-readahead-fedora - ../init.d/stop-readahead-fedora* lrwxrwxrwx 1 root root 25 2009-10-25 15:26 /etc/rcS.d/S02early-readahead - ../init.d/early-readahead* lrwxrwxrwx 1 root root 25 2010-03-26 10:03 /etc/rcS.d/S26later-readahead - ../init.d/later-readahead* * cat /proc/mounts It's pretty long, here's the top (all the normal fs's) rootfs / rootfs rw 0 0 none /sys sysfs rw,nosuid,nodev,noexec,relatime 0 0 none /proc proc rw,nosuid,nodev,noexec,relatime 0 0 udev /dev tmpfs rw,relatime,size=10240k,mode=755 0 0 /dev/disk/by-uuid/f27c45c6-187d-4bc8-a7b6-3c20315992a4 / xfs rw,relatime,noquota 0 0 tmpfs /lib/init/rw tmpfs rw,nosuid,relatime,mode=755 0 0 tmpfs /dev/shm tmpfs rw,nosuid,nodev,relatime 0 0 devpts /dev/pts devpts rw,nosuid,noexec,relatime,gid=5,mode=620,ptmxmode=000 0 0 tmpfs /tmp tmpfs rw,relatime 0 0 /dev/mapper/vg1-usr /usr xfs rw,relatime,attr2,logbufs=8,noquota 0 0 /dev/mapper/vg1-var /var xfs rw,relatime,logbufs=8,noquota 0 0 Mike Stone -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
Package: readahead-fedora Version: 2:1.5.4-5 Severity: normal as far as I can tell from looking at the scripts, there should be an early.sorted in /etc/readahead.d as well as a later.sorted. (My system has seperate /usr and /var partitions.) If I run a profile boot, only the early.sorted is created (and as expected it doesn't contain anything from the later stages of the boot process). Am I missing a config parameter or something? -- System Information: Debian Release: squeeze/sid APT prefers unstable APT policy: (500, 'unstable'), (1, 'experimental') Architecture: i386 (i686) Kernel: Linux 2.6.32-3-686 (SMP w/2 CPU cores) Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/bash Versions of packages readahead-fedora depends on: ii dpkg 1.15.5.6Debian package management system ii e2fslibs 1.41.11-1 ext2/ext3/ext4 file system librari ii libaudit01.7.13-1+b1 Dynamic library for security audit ii libblkid12.16.2-0block device id library ii libc62.10.2-6Embedded GNU C Library: Shared lib readahead-fedora recommends no packages. readahead-fedora suggests no packages. -- no debconf information -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#575537: readahead-fedora: doesn't create later.sorted
Hi, Thanks for your report. On 26 March 2010 10:54, Michael Stone mst...@debian.org wrote: Package: readahead-fedora Version: 2:1.5.4-5 Severity: normal as far as I can tell from looking at the scripts, there should be an early.sorted in /etc/readahead.d as well as a later.sorted. (My system has seperate /usr and /var partitions.) If I run a profile boot, only the early.sorted is created (and as expected it doesn't contain anything from the later stages of the boot process). Am I missing a config parameter or something? No, everything should be done automatically. Could you please provide the following information: * Non-comment contents of /lib/init/rw/readahead The output of: * ls -l /etc/readahead.d * sh -x /etc/cron.daily/readahead * ls -l /etc/readahead.d (after running the cron script by hand) * ls -l /etc/rc*.d/*readahead* * cat /proc/mounts That might give me an idea of what's going on. Cheers, -- Raphael Geissert - Debian Developer www.debian.org - get.debian.net -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org