Re: general/7570: under heavy traffic, root process dies with 'LOCK_EX: Error getting accept lock. Exiting!'

[Pekka Savola <pe...@netcore.fi>]

On 16 Apr 2001 marc@apache.org wrote:

> [In order for any reply to be added to the PR database, you need]
> [to include <ap...@Apache.Org> in the Cc line and make sure the]
> [subject line starts with the report component and number, with ]
> [or without any 'Re:' prefixes (such as "general/1098:" or      ]
> ["Re: general/1098:").  If the subject doesn't match this       ]
> [pattern, your message will be misfiled and ignored.  The       ]
> ["apbugs" address is not added to the Cc line of messages from  ]
> [the database automatically because of the potential for mail   ]
> [loops.  If you do not include this Cc, your reply may be ig-   ]
> [nored unless you are responding to an explicit request from a  ]
> [developer.  Reply only with text; DO NOT SEND ATTACHMENTS!     ]
>
>
> Synopsis: under heavy traffic, root process dies with 'LOCK_EX: Error getting accept lock. Exiting!'
>
> State-Changed-From-To: open-feedback
> State-Changed-By: marc
> State-Changed-When: Mon Apr 16 09:56:36 PDT 2001
> State-Changed-Why:
> Are you sure you simply aren't running the box out of file descriptors with your log resolver?  It doesn't seem likely that this problem is related very much to Apache; it normally works fine, then when you run some resource intensive other process, it has trouble getting resources.  The obvious conclusion would be that this other program is using up "too many" of some type of resource for your particular kernel config, thereby causing problems for other processes.

This also occurred when we moved to a new high-performance server.  The
log resolving didn't occur there yet.

When I changed apache to use USE_FCNTL_SERIALIZED_ACCEPT, it has now run
with 500-700 processes around the clock without any problems for two
weeks.

It appears that FCNTL may work better than FLOCK at least for FreeBSD 4.3.