[Server-devel] anaconda deletes /fsckoptions on F9 based XS

[Martin Langhoff]

On Thu, Sep 18, 2008 at 1:46 PM, Joshua N Pritikin <jpritikin at pobox.com> wrote:
> I have run servers like this for a few years with ext3. I was surprised
> how well it worked. I never got anything resembling file system
> corruption. ext3 worked like a charm.

Good to hear!

> What tended to fail was the RAID1
> set up. For some reason, the volumes got unsync'd and had to be
> reassembled manually. But after issuing 'mdadm -a /dev/md0 /dev/foo' by
> hand, then it worked fine.

Interesting. We aren't pushing for RAID (for various reasons) but some
local teams may want to run it. Is there a good, safe way to put this
in an init script so that it auto-fixes itself? Assuming that the
mdadm command be run "online", can you help me craft an init script
that

 - checks if any devices are software raid
 - checks for unsync'd state if possible
 - runs the fixup

...?

> The only other problem was actual hard drive
> failure. But there is nothing we can do about that anyway.

yeah - that's the prob with raid1. Bad power, mgfg problems and heat
(the main sources of trouble) will be identical for both drives
involved, so it's likely they'll die together.

cheers,



m
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff