[Systems] sunjammer.sugarlabs.org downtime - Wed Dec 30 9am-10am New_York time
tomeu at sugarlabs.org
Tue Dec 29 11:25:53 EST 2009
On Tue, Dec 29, 2009 at 17:23, Daniel Clark <dclark at pobox.com> wrote:
> The machine that hosts sunjammer.sugarlabs.org has several failed disks
> (we use triple-redundancy raid1, so all of the sunjammer.sugarlabs.org
> data is still on a "normal", double-redundancy raid1).
> I am tentatively planning on replacing them tomorrow morning, Wed Dec 30
> 9am-10am New_York time.
> Normally this would not require downtime, however unfortunately one of
> the dead disks is a 8GB SLC SSD that is may not be hot-swappable (used
> for boot, not in a pull-out chassis).
> The particular model of SSD we are using seems prone to early death
> (when we got the two of them, one was DOA), so we are going to just
> replace both of them with new 16GB SSDs from a different manufacturer.
> If there is time during the maintenance window, we are going to relocate
> all filesystem journals to the SSD RAID1, which some people have
> reported as giving significant performance improvements .
> I realize the timing could be better with Bernie being on vacation, but
> this work really needs to get done; if someone else who has root on
> sunjammer could be available on #fsfsys (freenode) during the
> maintenance window that would be useful.
I think I can be around, though I don't have much admin knowledge.
> I'll confirm this timing via a post to the fsfstatus identi.ca group .
>  Improving Ext3 performance by placing the journal on a Flash Disk
>  Monitor FSF and GNU Server Status (fsfstatus) group
> Daniel JB Clark | Sys Admin, Free Software Foundation
> pobox.com/~dclark | http://www.fsf.org/about/staff#danny
> Systems mailing list
> Systems at lists.sugarlabs.org
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
More information about the Systems