Apache 1.2.4: error_log: fcntl: F_SETLKW: No record locks available ????

Apache 1.2.4: error_log: fcntl: F_SETLKW: No record locks available ????

Post by Edwin Zuide » Wed, 19 Nov 1997 04:00:00



Recently I started an apache webserver on a SUN Sparcstation 5.
After running for several days, the error log starts generating
the following errors:

[Wed Oct 22 10:10:35 1997] fcntl: F_SETLKW: No record locks available
[Wed Oct 22 10:10:35 1997] - Error getting accept lock. Exiting!

and

[Mon Nov  3 15:16:04 1997] fcntl: F_SETLKW: Stale NFS file handle
[Mon Nov  3 15:16:04 1997] - Error getting accept lock. Exiting!

Then the server also starts to keep 'hanging' on the client message
"host contacted, waiting for reply..."

What is wrong?
File permissions? The page looks good, nothing is missing...

Thanks,

Edwin

_________________________________________
Edwin Zuidema, System Engineer
Hollandse Signaalapparaten B.V.
PO Box 42 7550 GD Hengelo The Netherlands
telephone: +31-74-248-2287


 
 
 

Apache 1.2.4: error_log: fcntl: F_SETLKW: No record locks available ????

Post by Marc Slemk » Wed, 19 Nov 1997 04:00:00



Quote:>Recently I started an apache webserver on a SUN Sparcstation 5.
>After running for several days, the error log starts generating
>the following errors:

>[Wed Oct 22 10:10:35 1997] fcntl: F_SETLKW: No record locks available
>[Wed Oct 22 10:10:35 1997] - Error getting accept lock. Exiting!
>and
>[Mon Nov  3 15:16:04 1997] fcntl: F_SETLKW: Stale NFS file handle
>[Mon Nov  3 15:16:04 1997] - Error getting accept lock. Exiting!
>Then the server also starts to keep 'hanging' on the client message
>"host contacted, waiting for reply..."
>What is wrong?
>File permissions? The page looks good, nothing is missing...

You don't mention what OS you are running, but I will guess Solaris.

Is the ServerRoot NFS mounted?  If so, add a "LockFile /var/tmp/accept.lock"
(or some other local path) directive to your config file and restart
Apache.

 
 
 

Apache 1.2.4: error_log: fcntl: F_SETLKW: No record locks available ????

Post by Nick Ke » Thu, 20 Nov 1997 04:00:00



> [Mon Nov  3 15:16:04 1997] fcntl: F_SETLKW: Stale NFS file handle
> [Mon Nov  3 15:16:04 1997] - Error getting accept lock. Exiting!

You have a deathwish?  Why is your webserver running from an NFS-
mounted disc?  I guess you're relying on automount here: does it
work for _any_ daemon process?

Nick Kew.

 
 
 

1. Record locks not available (Solaris+Apache)

Hi!

I'm having problems when running  Apache web server on my Ultras.
A while after the server starts running, Apache fails logging that
"FCNTL. No record locks available"
I have two twin machines, running identical Apache version, with two
different webs. Both of them fails the same.
Another two twin machines (identical to the previous two) runs a web
identical to one of the failing ones, and get no failures at all.

Does anybody know any tip I can search in order to fix the problem?

(I would thank email CC if possible)

Lots of thanks in advance!

_
Jorge Uriarte Aretxaga

2. proxy route gateway ip masquerading ip chains ?

3. vsftpd setup

4. Apache SSL: No records lock available

5. IIvx booting problems (fwd)

6. Problems with MMAP and/or FCNTL for record locking

7. Reading is much more interesting than TV (0194/1708)

8. Linux' fcntl record locking is broken

9. problems with record locking and fcntl()

10. Back to fcntl record locking with nfs and multiple descriptors

11. record locking with fcntl on NFS

12. File/Record Locking w/ fcntl()