Topic: Server running slow - log errors including Connection reset by peer, Broken pipe

[vaf]

Hi all

I logged a bug last Friday, but still got nothing back, so I thought I'd ask the forums.

Last Friday my SME Server 7.5.1 (new hardware C2D, 2GB DDR2, 5x2TB RAID) box went ridiculously slow all of a sudden. A reboot got it running again, however I want to avoid the issue happening. Anyway it happened again, and again a reboot "fixed" it for now, but we can't reboot the server every few days - and it happens without warning, or obvious reason.

You can see the issue and relevant outputs in the bug report here: http://bugs.contribs.org/show_bug.cgi?id=6336

Our SME server 7.5.1 machine had a fit overnight. I have been trying to address
an issue as per http://bugs.contribs.org/show_bug.cgi?id=6282. I have been
suggested to downgrade Samba (not sure how to do this - please let me know how
if you think I should).

Both file and server access became ridiculously slow somewhere overnight and
this morning it took about 5 mins to open 50kb file. The overnight backup found
file names but failed to copy - error log has cp: cannot access
'/job_001//XXX/./xxx...': Input/output error for every file and putty session
was also very slow with a couple of timeouts trying to get output attached. The
log file shows Error was Transport endpoint is not connected and Reset by Peer
errors.

I updated to 7.5.1 two weeks ago, and didn't do any changes or upgrades between
then and when the issues came up.

I did an upgrade with gave a new upgrade (can't remember which one - new log
file should show) and after reebooting the server has sped up.

The output.txt is pre-reboot output of
/sbin/e-smith/audittools/newrpms
/sbin/e-smith/audittools/templates
top

Issue is I need to prevent this from happening again.

Please help!

[Stefano]

(new hardware C2D, 2GB DDR2, 5x2TB RAID)

5x2TB.. hw o sw raid? what kind of raid?

[vaf]

Software raid. From the logs I don't think that it is a raid issue, but I am open to suggestions! Thanks

[piran]

open to suggestions!

Is your box supported by an appropriate UPS?
Please add the dmesg log file as an attachment
to your existing bugzilla entry. Five 2TB drives in
a SME software RAID take an awfully long time to
(re)sync, during which there can be more latency.
What's loading look like in HTOP?

[cactus]

I logged a bug last Friday, but still got nothing back, so I thought I'd ask the forums.

Please don't. This will scatter information. To all, please direct all information to http://bugs.contribs.org/show_bug.cgi?id=6282 .