[plug] Downtime

[Timothy White]

On 11/23/06, Cameron Patrick <cameron at patrick.wattle.id.au> wrote:
> Timothy White wrote:
>
> > We apologise for the downtime of the plug servers last night. I'm
> > still not sure what happened, as we were in the process of trying to
> > fix mod_perl, and we even lost our serial console connection.
>
> I suspect it was a hardware problem with shameless (unless you were
> doing something as root that was likely to hang the machine? the kernel
> we have on there is sensitive to ifup/ifdown-ing interfaces in the wrong
> order).

Well, looking at the logs, it just died. Nothing in there about a
shutdown or anything. And as it didn't even respond to the serial
console, it was really dead. The most we did as root, was restarting
apache!

>
> > On that note, any budding Debianites who wish to backport
> > libapache2-mod-perl2 to sarge, please put your hand up, as I've failed
> > a simple attempt at it!
>
> Does the version in etch fix things?  We're near enough to a release we
> could probably upgrade the PLUG machines now.

Well, if your willing to upgrade to etch.... I've not yet tried the
etch version, as it would require pulling in apache2.2 as well as a
newer perl.

>
> > p.s. I'm not sure who reset it
>
> Trent Lloyd, who works for Arachnet^W Host Away.  I grabbed him on IRC
> last night.

I suspected it was him. I was about to call him at Host Away when I
noticed that the machine had just come up! Thanks Trent!

Also, if anyone has a simple serial circuit, that can "short" a
circuit, or worst case, run a relay...
It would be nice to build a simple serial dongle, that goes on the
reset pins of shameless's motherboard, for completely remote reboot's!
(Although, it helps if spark doesn't hang for so long on NFS errors,
might have to look at that, timeo decrease?)

Tim
-- 
Linux Counter user #273956
Don't email joeblogs at scouts.org.au