quick question on sshd_log

368 views
Skip to first unread message

Christian MICHON

unread,
May 10, 2010, 10:33:19 AM5/10/10
to Repo and Gerrit Discussion
Hi,

on lines with git-upload-pack info, what is the meaning of 128 on the
last field ?
(I assume 0 is ok, but I'd like to know what is the meaning of other codes).

example:

[2010-05-10 15:19:26,002 +0200] xxxxxxx yyyyyy a/125 'git-upload-pack
'\''blabla.git'\''' 28368ms 1456452ms 128

Thanks in advance

--
Christian
--
http://detaolb.sourceforge.net/, a linux distribution for Qemu with Git inside !

--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

Shawn Pearce

unread,
May 10, 2010, 11:00:12 AM5/10/10
to Christian MICHON, Repo and Gerrit Discussion
Christian MICHON <christia...@gmail.com> wrote:
> on lines with git-upload-pack info, what is the meaning of 128 on the
> last field ?

Its the exit status of the task.

> (I assume 0 is ok, but I'd like to know what is the meaning of other codes).

Right, 0 is OK. Anything else is "not OK".

A status of 128 means an exception was caught internally, should
have been logged to the error_log, and the remote command ended
unexpectedly. 128 happens to also be the exit status of a git-core
command which has called its internal die() handler and terminated
early for any number of reasons.

I would check your error_log around the same time. There should
be an exception trace correlating to this log event.

> [2010-05-10 15:19:26,002 +0200] xxxxxxx yyyyyy a/125 'git-upload-pack
> '\''blabla.git'\''' 28368ms 1456452ms 128

Christian MICHON

unread,
May 10, 2010, 11:52:12 AM5/10/10
to Shawn Pearce, Repo and Gerrit Discussion
Thanks for the quick hint. More questions below

On Mon, May 10, 2010 at 5:00 PM, Shawn Pearce <s...@google.com> wrote:
>> (I assume 0 is ok, but I'd like to know what is the meaning of other codes).
>
> Right, 0 is OK.  Anything else is "not OK".
>
> A status of 128 means an exception was caught internally, should
> have been logged to the error_log, and the remote command ended
> unexpectedly.  128 happens to also be the exit status of a git-core
> command which has called its internal die() handler and terminated
> early for any number of reasons.
>
> I would check your error_log around the same time.  There should
> be an exception trace correlating to this log event.

looks like this:
ERROR com.google.gerrit.sshd.BaseCommand : Internal server error (user
yyyyy account 125) during git-upload-pack '/blahblah.git'

and most important:
java.lang.OutOfMemoryError: GC overhead limit exceeded

Does this mean the server is not configured according to the size of
data (blahblah is actually equivalent to a linux kernel) ?

Which config file should I look at ? (I'm not the admin of this
server, but slow down were seen, I'd like to debug this)

Thanks again

--
Christian
--
http://detaolb.sourceforge.net/, a linux distribution for Qemu with Git inside !

Shawn Pearce

unread,
May 10, 2010, 12:07:31 PM5/10/10
to Christian MICHON, Repo and Gerrit Discussion
Christian MICHON <christia...@gmail.com> wrote:
> On Mon, May 10, 2010 at 5:00 PM, Shawn Pearce <s...@google.com> wrote:
> >> (I assume 0 is ok, but I'd like to know what is the meaning of other codes).
> >
> > A status of 128 means an exception was caught internally
>
> looks like this:
> ERROR com.google.gerrit.sshd.BaseCommand : Internal server error (user
> yyyyy account 125) during git-upload-pack '/blahblah.git'
>
> and most important:
> java.lang.OutOfMemoryError: GC overhead limit exceeded

Yup. That's probably the issue.

> Does this mean the server is not configured according to the size of
> data (blahblah is actually equivalent to a linux kernel) ?

Yes. The server's JVM heap isn't big enough to handle the workload
being thrown at it. This is set by the container.heapLimit variable
in gerrit.config, if the admin is starting gerrit through gerrit.sh.
Reply all
Reply to author
Forward
0 new messages