Re: zfs meltdown on sun blade 150
- From: kangcool <kangcool2002@xxxxxxxxxxx>
- Date: Tue, 03 Jul 2007 17:26:47 -0700
There is a patch for this.
if you have already got it then we got a new problem :)
Ram should be ok, from the sound of it the patch should solve it. If
not post to the zfs forum on the opensolaris site, thats where most of
the zfs people lurk.
Tim Bradshaw wrote:
On Jul 3, 10:57 pm, d...@xxxxxxxx wrote:
I've been trying out a zfs mirrored filesystem on a sun blade 150 with
1 gbyte of memory,
and the performance has been disastrous. I made the mistake of
putting one mirror on a
slice of a disk which also has the root filesystem, so that's part of
the problem. Perhaps
the 1 Gbyte of memory is another. But maybe this is just too small a
machine to use
zfs?
There is a known problem with the IDE controller Sun used (use?
presumably anything desktopy now would be SATA) and ZFS.
I hear rumours (in this group, last few weeks) of a patch for this.
Failing that one horrid workaround is to create simple SVM devices and
use them to build the ZFS pool. Because the SVM fevices map 1-1 onto
the underlying slices you can then move to using them by exporting &
reimporting the pool when a fix appears (if it has not already).
--tim
.
- References:
- zfs meltdown on sun blade 150
- From: danq
- Re: zfs meltdown on sun blade 150
- From: Tim Bradshaw
- zfs meltdown on sun blade 150
- Prev by Date: Re: zfs meltdown on sun blade 150
- Next by Date: Mailx with attachments
- Previous by thread: Re: zfs meltdown on sun blade 150
- Next by thread: Mailx with attachments
- Index(es):
Relevant Pages
|
