System halted

[Ruben Cardenal]

Hi,

From the syslog.log file:

Jan 23 14:39:36 grecopr2 inetd[553]: login/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 cmlvmd: Failed to accept connections from commands: No buffer space available
Jan 23 14:39:36 grecopr2 cmlvmd: CLVMD exiting
Jan 23 14:39:36 grecopr2 inetd[553]: printer/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: discard/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: exec/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: kshell/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 cmcld: Unable to accept a connection: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: hacl-cfg/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: ident/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: telnet/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: ftp/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: hacl-probe/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: dtspc/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: klogin/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: daytime/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 remshd[16887]: Connection from 0.0.0.0 on illegal port
Jan 23 14:39:36 grecopr2 inetd[553]: echo/tcp: accept: No buffer space available
Jan 23 14:39:42 grecopr2 cmcld: Processing exit status for service cmlvmd
Jan 23 14:39:42 grecopr2 cmcld: Service cmlvmd terminated due to an exit(1).
Jan 23 14:39:42 grecopr2 cmcld: Halting grecopr2 to preserve data integrity
Jan 23 14:39:42 grecopr2 cmcld: Reason: LVM daemon failed
Jan 23 14:40:04 grecopr2 vmunix:
Jan 23 14:40:04 grecopr2 vmunix: sync'ing disks (0 buffers to flush):
Jan 23 14:40:04 grecopr2 vmunix: 0 buffers not flushed
Jan 23 14:40:04 grecopr2 vmunix: 0 buffers still dirty

And system rebooted. Why?

[Krishna Prasad]

It's hard to tell from the syslog if these messages caused the reboot or where caused by something else?

Did you have HP look at the /var/tombstones/ts99 file? Check the time stamp on these file and see if it close to the time of the reboot. It is possible it will tell you more.

Also, if you have any files in /var/adm/crash that you can run Q4 against.

Hope this helps

[Jeff Machols]

take a look at this thread

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0xe0f8a12d6d27d5118fef0090279cd0f9,00.html

[Steve Steel]

Hi

There is a known issue where ServiceGuard
aborts as result of the accept() system call returning an ENOBUFS,
"No buffer space available".

This issue where the Serviceguard daemon aborts as a result of the
accept() returning an ENOBUFS is a known problem fixed in Serviceguard
versions A.11.07 and beyond where a check has been added for the
EINVAL and ENOBUFS conditions.
nown problem

Which version do you have


Steve Steel

[Paula J Frazer-Campbell]

Ruben

Please attach your ts99 file.

Paula

[Frederic Sevestre]

Hi,

The reboot was caused by service guard.
The cmlvmd was not able to talk with the other node so it exit and cause a TOC.
I don't know the reason of the "no buffer space".

Fr??d??ric

[melvyn burnard]

This has nothing to do with tombstones etc.
The reason the system rebooted is in the log file:

>Jan 23 14:39:36 grecopr2 cmlvmd: CLVMD exiting
and you appear to have run out of memory for networking:

Jan 23 14:39:36 grecopr2 cmcld: Unable to accept a connection: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: hacl-cfg/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: hacl-probe/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: dtspc/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: klogin/tcp: accept: No buffer space available
Jan 23 14:39:36 grecopr2 inetd[553]: daytime/tcp: accept: No buffer space available

Hence CMCLD decides cmlvmd is no longer working, and then perfoms and orderly shutdown of the system:

Jan 23 14:39:42 grecopr2 cmcld: Processing exit status for service cmlvmd
Jan 23 14:39:42 grecopr2 cmcld: Service cmlvmd terminated due to an exit(1).
Jan 23 14:39:42 grecopr2 cmcld: Halting grecopr2 to preserve data integrity
Jan 23 14:39:42 grecopr2 cmcld: Reason: LVM daemon failed
Jan 23 14:40:04 grecopr2 vmunix:
Jan 23 14:40:04 grecopr2 vmunix: sync'ing disks (0 buffers to flush):
Jan 23 14:40:04 grecopr2 vmunix: 0 buffers not flushed
Jan 23 14:40:04 grecopr2 vmunix: 0 buffers still dirty


So that is why the system rebooted ServiceGuard told it to.
As to why ytoiu ran out of memory? could be memory leak or you simply have insufficient memory.

[Steve Steel]

Hi

There is a known problem.

See my previous addition.


Steve Steel