CVS SERVER OUT OF SPACE?

[Oliver Stieber]

Hi,
It looks like cvs server has run out of diskspace.

cvs update reports.

cvs server: Updating .
cvs server: failed to create lock directory for
`/home/wine/wine' (/home/wine/wine/#cvs.lock): No
space left on device
cvs server: failed to obtain dir lock in repository
`/home/wine/wine'
cvs [server aborted]: read lock failed - giving up

Send instant messages to your online friends http://uk.messenger.yahoo.com

[Jeremy Newman]

Sorry about that. All fixed now.

On Tue, 2005-03-01 at 17:04 +0000, Oliver Stieber wrote:
> Hi,
> It looks like cvs server has run out of diskspace.
>
> cvs update reports.
>
> cvs server: Updating .
> cvs server: failed to create lock directory for
> `/home/wine/wine' (/home/wine/wine/#cvs.lock): No
> space left on device
> cvs server: failed to obtain dir lock in repository
> `/home/wine/wine'
> cvs [server aborted]: read lock failed - giving up
>
> Send instant messages to your online friends http://uk.messenger.yahoo.com

[Andreas Mohr]

Hi,

On Tue, Mar 01, 2005 at 03:00:50PM -0600, Jeremy Newman wrote:
> Sorry about that. All fixed now.

The main question is: is it "fixed" or is it "FIXED"?
There's a difference, ya know...

"FIXED" meaning that this will occur never again due to an SNMP alert
infrastructure (or something equally functional) being in place...
(which I'd have assumed to be there already for WineHQ, given the relative
importance of this server)

Or maybe it simply was a temporary hickup/user error despite an alert
infrastructure being in place already? (which thus must have been
non-operational / ignored this time)

Greetings,

Andreas Mohr

[Jeremy White]

This was a CodeWeavers related function that is taking more
disk space on our main server that we anticipated.  All we
did was clean up some space and warn folks within CodeWeavers that this can
happen; that should suffice until we make the transition to the
new set of servers that Jer has built up.

The new servers are for CW; WineHQ gets to stay on the older server, but doesn't have to
share it anymore.  This is something Jeremy has been desparate to do fo years.
His cheapskate boss has finally relented <grin>).

We meet with the colo people tomorrow to plan the extra half rack we need
and the switch will likely happen in the next few months.

Cheers,

Jer

Andreas Mohr wrote:
> Hi,
>
> On Tue, Mar 01, 2005 at 03:00:50PM -0600, Jeremy Newman wrote:
> > Sorry about that. All fixed now.
>
> The main question is: is it "fixed" or is it "FIXED"?
> There's a difference, ya know...
>
> "FIXED" meaning that this will occur never again due to an SNMP alert
> infrastructure (or something equally functional) being in place...
> (which I'd have assumed to be there already for WineHQ, given the relative
> importance of this server)
>
> Or maybe it simply was a temporary hickup/user error despite an alert
> infrastructure being in place already? (which thus must have been
> non-operational / ignored this time)
>
> Greetings,
>
> Andreas Mohr

[Jeremy Newman]

On Wed, 2005-03-02 at 09:18 +0100, Andreas Mohr wrote:
> "FIXED" meaning that this will occur never again due to an SNMP alert
> infrastructure (or something equally functional) being in place...
> (which I'd have assumed to be there already for WineHQ, given the relative
> importance of this server)

Yes, I have server monitoring software in place. It warns me via email
when the server breaks certain thresholds. I.e. disk space, process
count, zombie count, user count, load average, etc.

This happened because we had someone uploading us a linux distro DVD to
our upload space on that server. I have remedied this problem by letting
our users know to use one of our internal servers, not the web server to
upload large files such as this.

[Andreas Mohr]

Hi,

On Wed, Mar 02, 2005 at 08:30:17AM -0600, Jeremy Newman wrote:
> On Wed, 2005-03-02 at 09:18 +0100, Andreas Mohr wrote:
> > "FIXED" meaning that this will occur never again due to an SNMP alert
> > infrastructure (or something equally functional) being in place...
> > (which I'd have assumed to be there already for WineHQ, given the relative
> > importance of this server)
>
> Yes, I have server monitoring software in place. It warns me via email
> when the server breaks certain thresholds. I.e. disk space, process
> count, zombie count, user count, load average, etc.

Good guy :)
> This happened because we had someone uploading us a linux distro DVD to
> our upload space on that server. I have remedied this problem by letting
> our users know to use one of our internal servers, not the web server to
> upload large files such as this.

The usual super-catastrophic events, it seems...
(which probably could have been avoided here by having a separate partition
or quota for the upload directory, but some critical things to be protected
always tend to escape...)

Andreas Mohr