Hello,
Where exactly do you have the message "too many open files" ?
To verify this issue, if you can isolate the process that displays this
message, use the following
command to verify the number of open files
pfiles $(pgrep nsrexecd) <-- change nsrexecd by save or savefs or your backup
command.
Updating the file /etc/system will change the maximum number of open for the
overall system, which
may be is not the best practice.
The second option is to modify the startup script of the Networker daemons and
adding at the top
of the script the line "ulimit -n 1024".
By default a process is started with a maximum of 256 file descriptors.
After changing the startup script of networker /etc/init.d/networker, you'll
have to restart the
daemons.
( cd / && sh /etc/init.d/networker stop && sh /etc/init.d/networker start)
The client daemons nsrexecd will be restart, you can verify that the file
descriptor limit has
been updated by typing
pfiles $(pgrep nsrexecd)
As all the backup commands (save, savefs, nsrnmo,...) are initated by the
nsrexecd daemons, then
will use the same file descriptot limit.
The example here will increase the number of open files from 256 to 1024, if
you want more you'll
have to update the file /etc/system first.
I've never seen such an issue before...
Regards.
--- Andrew Harkin <andrew.harkin_networker AT MCKENNAN DOT ORG> a écrit :
> This parameter is changed in /etc/system on a sun 2.8 box, which then
> requires a reboot to apply it. In my experience, I've generally had to bump
> some of those default system parameters up in Solaris, as any number of
> programs I've worked with seem to run into those and similar issues.
>
> As far as what you should set it to, I would do some research on it.
> Sunmanagers.org's listmail would be a good spot to check, I'm sure someone's
> asked for recommendations on it before, and there's a lot of very
> knowledgeable people in that group.
>
> Andrew Harkin
>
>
>
> -----Original Message-----
> From: Chad Smykay [mailto:csmykay AT RACKSPACE DOT COM]
> Sent: Wednesday, January 05, 2005 3:16 PM
> To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
> Subject: Re: [Networker] To many open files
>
> Ken,
>
> When you have "Too many open files" errors in ANY of your logs this is
> usually indicative of an OS problem. I am not that familiar with Solaris so
> I am not sure how the underlying kernel allows for more or less processes
> along with files to be opened and closed for such processes. You really
> should check with someone that knows SUN fairly well that can look at this
> issue. But I can tell you that it is not a Legato issue.
>
> Regards,
>
> Chad Smykay, RHCE, LCNA
> Systems Storage Administrator
> Rackspace Managed Hosting (TM) - The Managed Hosting Specialist (TM)
>
> ____________________
> Confidentiality Notice: This e-mail message, including any attachments, is
> for the sole use of the intended recipient(s) and may contain confidential
> and privileged information. Any unauthorized review, use, disclosure, or
> distribution is prohibited. If you are not the intended recipient, please
> contact the sender by reply e-mail and destroy all copies of the original
> message.
>
> --
> Note: To sign off this list, send a "signoff networker" command via email
> to listserv AT listmail.temple DOT edu or visit the list's Web site at
> http://listmail.temple.edu/archives/networker.html where you can
> also view and post messages to the list. Questions regarding this list
> should be sent to stan AT temple DOT edu
> =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
>
Découvrez le nouveau Yahoo! Mail : 250 Mo d'espace de stockage pour vos mails !
Créez votre Yahoo! Mail sur http://fr.mail.yahoo.com/
--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list. Questions regarding this list
should be sent to stan AT temple DOT edu
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=
|
