[Operators] jabber.org stability (was: Re: requiring channel encryption)

Mickaël Rémond mickael.remond at process-one.net
Fri May 2 15:04:44 CDT 2008


Hello,

Mnesia is actually transactionnal (and clustered).
Not sure what happened on the server at this time.
The best approach is anyway to set up disk space monitor to anticipate  
on those situations.

-- 
Mickaël Remond

Le 2 mai 08 à 21:10, "Kevin Smith" <kevin at kismith.co.uk> a écrit :

> On Fri, May 2, 2008 at 8:04 PM, Tomasz Sterna <tomek at xiaoka.com>  
> wrote:
>>> All databases do not like when they cannot write on disk.
>> I don't know others, but PostgreSQL is handling it very graceful.
>> The write operations fail (transactions are rolled back) but besides
>> that the server is running fine.
>> As you free some space, situation backs to normal and all new  
>> operations
>> work fine.
>
> Yes, Postgres behaved fine when we ran out of space - it's mnesia that
> corrupted all the data. I guess it's non-transactional, or such.
>
> /K
>


More information about the Operators mailing list