On Tue, 25 Sep 2007, Oleg Drokin wrote:
> Hello!
>
> On Tue, Sep 25, 2007 at 03:21:55PM +0200, Niklas Edmundsson wrote:
>> We managed to trigger another intriguing lustre failure by simply
>> rebooting our two OSS's (which has two OST's each), and then mounting
>> the OST's on both OSS's in parallell (ie mount ost1 ; mount ost2 on
>> both OSS's at the same time). The filesystem was mounted by 191
>> clients, of which twenty-ish hade been recently active. Interconnect
>> used is standard tcp over GigE.
>
> First of all you need to have a patch from bug 13614 if yo don't have it yet.
> https://bugzilla.lustre.org/attachment.cgi?id=12679
Nope, I didn't have that one.
I'll apply it and see if it improves the situation.
/Nikke
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se | [EMAIL PROTECTED]
---------------------------------------------------------------------------
Iraqi Bingo: B-52... F-16... M-1... F-18... F-117...
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
_______________________________________________
Lustre-discuss mailing list
[email protected]
https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
