dump after every reboot: is this normal?

[ponetguy2]

Every time I reboot our solaris 9 (SunOS 5.9 Generic_118558-19 sun4u sparc SUNW,Sun-Fire-480R) box,
I get the messages below:
# dmesg | grep dump
Mar 24 12:39:55 hostname savecore: [ID 346688 auth.error] initial dump header corrupt
Mar 24 12:39:55 hostname genunix: [ID 454863 kern.info] dump on /dev/dsk/c1t0d0s1 size 700 MB
Mar 24 12:48:08 hostname genunix: [ID 454863 kern.info] dump on /dev/md/dsk/d10 size 700 MB
Mar 24 12:48:22 hostname genunix: [ID 454863 kern.info] dump on /dev/dsk/c1t0d0s1 size 700 MB

Is this something I should be concerned about? Is this normal?

I checked our other solaris 9 machines and some have it and some don't.

 

[marrow]

who -b
system boot Feb 27 05:07
hostname $ grep dump /var/adm/m*

/var/adm/messages.Feb 27 05:03:52 hostname genunix: [ID 111219 kern.notice] dumping to /dev/md/dsk/d21, offset 429850624, content: kernel
/var/adm/messages.Feb 27 05:05:11 hostnmae genunix: [ID 409368 kern.notice] ^M100% done: 43013 pages dumped, compression ratio 2.90,
/var/adm/messages.Feb 27 05:05:11 hostname genunix: [ID 851671 kern.notice] dump succeeded
/var/adm/messages.Feb 27 05:07:19 hostname genunix: [ID 454863 kern.info] dump on /dev/md/dsk/d21 size 2049 MB


This is also Solaris9 O/S

 

[marrow]

Apologies, the above system is also an example of a system which crashed, normally we don't seem to have these messages unless there is a problem.

 

[huebs]

Use "savecore -d" to write out the dump disregarding the dump header valid flag.
Use dumpadm(1m) to configure your dump device, the default is to use the swap partition.

 

[elgrandeperro]

Are the devices it is listing your swap? Don't know why it lists the d10 meta and the raw disk twice.

When you reboot, do it from the console and make sure
it isn't writing to the dump device. If swap=dump device, I believe the dump image should be cleared, so I am curious why it does it every time.

How do you "reboot"?

gene

 

[ponetguy2]

I made some changes. I changed my dump device to /dev/md/dsk/d10. I did this by
doing a dumpad -d swap. I am running disksuite on this machine and Oracle. There seem to be
an issue with disksuite and dumpadm's dump device default. I should have done a dumpadm -f swap when
I initially mirrored this box. Oooops

hostname# dumpadm
Dump content: kernel pages
Dump device: /dev/md/dsk/d10 (swap)
Savecore directory: /var/crash/hostname
Savecore enabled: yes

I also renamed S75savecore to S22savecore. I found another forum where they posted that they were getting
the same messages as I am:

Mar 27 09:23:58 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

They claim that changing the startup sequence of S75savecore to S22savecore made the log above disapear.
I am a little skeptical with this solution. I think it is caused by a serial cable currently plugged in to
the machine. I will try to reboot this box with the serial cable unplugged and see if the log persist.
Hopefully the change I made will not hurt anything.

I'm still trying to figure out what is causing the dump during startup/reboot. I'm not doing anything special.
I do a shutdown -y -g0 -i6 to reboot the machine.

Here are the new messages:

Mar 24 12:48:23 hostname rdriver: [ID 400281 kern.notice] ID[RAIDarray.rdaemon.1001] RDAC Resolution Daemon locked in memory
Mar 24 12:48:27 hostname sshd[649]: [ID 800047 auth.info] Server listening on :: port xx.
Mar 24 12:48:28 hostname su: [ID 366847 auth.info] 'su oracleuser' succeeded for user on /dev/console
Mar 24 12:48:39 hostname last message repeated 1 time
Mar 24 12:49:00 hostname pseudo: [ID 129642 kern.info] pseudo-device: devinfo0
Mar 24 12:49:00 hostname genunix: [ID 936769 kern.info] devinfo0 is /pseudo/devinfo@0

Mar 24 12:49:41 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

Mar 24 12:49:59 hostname sshd[1096]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2

Mar 24 12:50:59 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

Mar 24 13:51:01 hostname sshd[2881]: [ID 800047 auth.info] Accepted password for oracleuser from 10.x1.1x.110 port xxx ssh2
Mar 24 13:51:09 hostname sshd[2895]: [ID 800047 auth.info] Accepted password for oracleuser from 10.x1.1x.110 port xxx ssh2

Mar 24 14:28:13 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

Mar 24 15:06:09 hostname sshd[5654]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2
Mar 24 16:07:52 hostname sshd[7196]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2

Mar 24 16:19:50 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

Mar 27 07:37:28 hostname sshd[4309]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.44 port xxx ssh2
Mar 27 08:33:24 hostname sshd[5393]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2
Mar 27 08:39:38 hostname sshd[5525]: [ID 800047 auth.info] Accepted password for oracleuser from 10.x1.1x.110 port xxx ssh2
Mar 27 08:39:45 hostname sshd[5540]: [ID 800047 auth.info] Accepted password for oracleuser from 10.x1.1x.110 port xxx ssh2
Mar 27 09:08:19 hostname sshd[6138]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2

Mar 27 09:23:58 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

Mar 27 09:25:14 hostname sshd[6791]: [ID 800047 auth.info] Accepted password for oracleuser from 10.x1.1x.110 port xxx ssh2
Mar 27 10:10:35 hostname sshd[7787]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2
Mar 27 10:33:08 hostname sshd[8160]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2
Mar 27 10:37:57 hostname sshd[8328]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2

Mar 27 10:38:09 hostname genunix: [ID 454863 kern.info] dump on /dev/md/dsk/d10 size 700 MB

Mar 27 11:38:29 hostname sshd[9776]: [ID 800047 auth.info] Accepted password for user from 10.x1.1x.112 port xxx ssh2

I will do another reboot this afternoon to see if the messages persist.

 

[ponetguy2]

I unplugged the serial connection from the machine and question and did a reboot. No dice. The messages still persist. I'm totally lost as to what is causing this messages. GRRRRRRRRRRR!!!

# cat /var/adm/messages | grep se_hdlc
Mar 27 15:51:22 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!
Mar 27 16:13:35 hostname se: [ID 427429 kern.warning] WARNING: se_hdlc: clone device must be attached before use!

# cat /var/adm/messages | grep dump
Mar 27 15:47:29 hostname genunix: [ID 454863 kern.info] dump on /dev/md/dsk/d10 size 700 MB
Mar 27 16:09:13 hostname genunix: [ID 454863 kern.info] dump on /dev/md/dsk/d10 size 700 MB

 

[huebs]

There's nothing wrong about the "... dump on ..." lines in the messages after doing a reboot.

 

[ponetguy2]

thanx huebs. i was'nt sure if the dump is normal during reboot. I just wish it is documented somewhere so I can show my boss.

 

[huebs]

It's just an information telling location and size of the dump device.

 

[ponetguy2]

thank you huebs. i'm just curious why some solaris boxes log this messages and some don't. i'm a little worried about this. plus i need to explain to my boss that this message is normal and i need to prove it via some type of documentation. i'm in a tough spot. i really appreciate your help.

this machine is possesed since we got hold of it. there is always something wrong with it. now i'm getting this messages:

Mar 28 15:15:10 hostname krtld: [ID 469452 kern.info] NOTICE: sf: 64-bit driver module not found

looks like i need to find a driver for my scsi device. plus we managed to get rid of se_hdlc messages. i disabled a startup script implemented by our database admin which got rid of the se_hdlc message. it was kicking off all types of network services process.

 

[huebs]

Sorry, I couldn't spot any reference to the kernel message regarding the dump device at docs.sun.com ... just to help you with your boss.
We are running Solaris 8 and 10 here and all recently rebootet machines show the log entry "<hostname> genunix: [ID 454863 kern.info] dump on <device> size <x>". If you can't find the entry on some of your boxes, maybe they have a long uptime (the logs are already rotated away) or they have a modified syslog.conf

In Solaris 8, the sf driver is in package SUNWluxd (32bit) and SUNWluxdx (64bit). Solaris 10 does not seem to have an 32bit sf driver, SUNWluxd just contains the sparcv9 binary. Try "pkgchk -v SUNWluxd SUNWluxdx".

 

[kHz]

These links describe your kernel runtime linker message:

http://unix.derkeiler.com/Mailing-Lists/SunManagers/2003-08/0248.html

http://www.sunmanagers.org/pipermail/sunmanagers/2003-September/025581.html

Don't know about the dump from syslog. Some of my servers have the message and some don't. It's only a kernel information message though, nothing to worry about. You could do a find for a core with size <N> or see if you have anything in /var/crash/<hostname>.

 

[ponetguy2]

i seem to be missing the sf driver w/64 bit support:

# pkginfo |grep -i lux
system SUNWeulux UTF-8 L10N For Language Environment User Files (64-bit)
system SUNWluxd Sun Enterprise Network Array sf Device Driver
system SUNWluxl Sun Enterprise Network Array socal Device Driver
system SUNWluxop Sun Enterprise Network Array firmware and utilities
system SUNWluxox Sun Enterprise Network Array libraries (64-bit)
system SUNWvolux Volume Management (Usr) (64-bit)

 

[ponetguy2]

thank you kHz. those links are very helpfull. i will install the packages and patches asap.

i need to find a way to prove to my boss that the dump during reboot is normal. i'll keep you guys posted