[Linerva-announce] Linerva reboot today

[Geoffrey Thomas]

Hi,

Linerva crashed this morning at around 9 AM and has been restarted. The 
crash is the same as we saw on Sunday; there are also indications that 
this may have been caused of an Athena-wide network blip. We've restarted 
Linerva and will look into exactly what happened here.

Apologies for the inconvenience; as always feel free to contact us at 
linerva at mit.edu if you have any questions.

-- 
Geoffrey Thomas
SIPB Linerva team
linerva at mit.edu

On Sun, 24 Jul 2011, Geoffrey Thomas wrote:

> Hi,
>
> From approximately 1 PM to 5 PM today, Linerva was offline due to an AFS 
> crash. This outage was much longer than usual because we upgraded to a newer 
> prerelease (pre7) of OpenAFS 1.6, which should bring several stability 
> improvements.
>
> We were planning to schedule downtime for Linerva later this week to upgrade 
> to OpenAFS 1.6.0pre7. In addition to having wanted to do so for some time, 
> this was made more urgent by increasing AFS instability problems over the 
> past week. (This reboot should resolve any "Connection timed out" errors you 
> may have seen in normal operation on Linerva; if you still see those, please 
> let us know.) Unfortunately, the AFS instability issues caused a kernel crash 
> at 1 PM today which we could not recover from, and we decided for the sake of 
> future stability to continue with the plan of building and upgrading OpenAFS 
> 1.6.0pre7 for Linerva, instead of rebooting quickly and scheduling further 
> downtime for later in the week.
>
> Linerva is back online now and should be stable. We apologize for the 
> inconvenience and the extended outage; if you have any questions, feel free 
> to contact us at linerva at mit.edu.
>
> -- 
> Geoffrey Thomas
> SIPB Linerva team
> linerva at mit.edu
>