Extremely slow reboot CentOS 4.3

[Lyttek]

I'm having a problem with a new server.

CentOS 4.3, Linux version 2.6.9-34.0.2.EL

Reboots take forever! Notice that in the log entries below, nearly 23 minutes elapse!

Aug 8 08:29:39 dedi125 syslog: klogd shutdown succeeded
Aug 8 08:29:39 dedi125 exiting on signal 15
Aug 8 08:52:57 dedi125 syslogd 1.4.1: restart.
Aug 8 08:52:57 dedi125 syslog: syslogd startup succeeded

The datacenter (crystaltech) is baffled, as am I. Any ideas?

(cPanel not installed yet)

 

[HostMerit]

Try:

shutdown -rf now

That will do a fast reboot and skip fsck on the reboot, that may be causing it. If it reboots fast this way, you willknow its fscking every time on reboot.


Also crystaltech wouldn't be baffled if someone took the time to reboot from console and watched it boot up with a crash cart...

 

[Lyttek]

Try:

shutdown -rf now

Well, that dropped it down to 13 minutes... an improvement, but still...

On another box, just compared, the pause between entries as mentioned above was all of 90 seconds.

 

[Lyttek]

Is there *anything* that jumps out as being a cause in these two bits:

Slow box:

# This file is edited by fstab-sync - see 'man fstab-sync' for details
LABEL=/                 /                       ext3    defaults        1 1
LABEL=/boot             /boot                   ext3    defaults        1 2
none                    /dev/pts                devpts  gid=5,mode=620  0 0
none                    /dev/shm                tmpfs   noexec,nosuid   0 0
none                    /proc                   proc    defaults        0 0
none                    /sys                    sysfs   defaults        0 0
LABEL=SWAP-hdc2         swap                    swap    defaults        0 0

Fast box:

# This file is edited by fstab-sync - see 'man fstab-sync' for details
/dev/VolGroup00/LogVol00 /                       ext3    defaults,usrquota     $
LABEL=/boot             /boot                   ext3    defaults        1 2
none                    /dev/pts                devpts  gid=5,mode=620  0 0
none                    /dev/shm                tmpfs   noexec,nosuid   0 0
none                    /proc                   proc    defaults        0 0
none                    /sys                    sysfs   defaults        0 0
/dev/VolGroup00/LogVol01 swap                    swap    defaults        0 0

/dev/hdc1 /backups ext3 defaults 1 2

/dev/hdd1 /home2 ext3 defaults,usrquota 1 2
/dev/hdb                /media/cdrecorder       auto    pamconsole,exec,noauto,$
/dev/fd0                /media/floppy           auto    pamconsole,exec,noauto,$

 

[rpsands]

You might want to check the server's memory addressing.

We've had some issues with Cent 4.3 and memory addressing on certain motherboards (P4-m800-mx). The server runs, but it takes 20-30 minutes to boot up and is amazingly sluggish.

The only fix, currently, is to modify the mtrr stuff in rc.sysinit. What you put in there depends on the precise amount of memory you have.

Ask them to compare /proc/mtrr on your server to /proc/mtrr on another server, and see whether there're any obvious differences.

On a normal functioning server it'll usually look something like this (believe 1mb of video ram);

reg00: base=0x00000000 ( 0MB), size=2048MB: write-back, count=1
reg01: base=0x7ff00000 (2047MB), size= 1MB: write-through, count=1


On a screwy cent+p4m800 server you'll have a lot more going on in there. The memory'll be broken up into small chunks and have like 5-6 or more entries in /proc/mtrr.

This may have something to do with it. From what I know of CT they only use uniform hardware though so I don't see why that would be the case.

Original article:
http://groups.google.com/group/linux.kernel/msg/83dd8aa7e785c8dc