Prev: Start program from boot, which needs library in non-standard place
Next: Trying to optimize PHP on a T5140
From: Richard B. Gilbert on 4 Feb 2010 19:14
David Kirkby wrote:
> On Feb 4, 5:55 pm, cindy <cindy.swearin...(a)sun.com> wrote:
>> On Feb 4, 10:48 am, cindy <cindy.swearin...(a)sun.com> wrote:
>>
>>
>>
>>> On Feb 4, 7:46 am, David Kirkby <drkir...(a)gmail.com> wrote:
>>>> I've hit a problem that a Google search shows others have had before -
>>>> trying to create a snapshot, one gets a message that the dataset is
>>>> busy.
>>>> # zfs snapshot rootpool2/ROOT/S10u7zfsB(a)today
>>>> cannot create snapshot 'rootpool2/ROOT/S10u7zfsB(a)today': dataset is
>>>> busy
>>>> The usual way around this is to unmount then remount the file system.
>>>> That is not always a practical option though. I can't possibly unmount
>>>> a file system that many are using. One of the benefits of ZFS is the
>>>> snapshots. Being told the dataset is busy is not a lot of use.
>>>> Is there a better solution, or a decent explanation of why this
>>>> happens?
>>>> Dave
>>> Hi Dave,
>>> One explanation of this problem was a bug, but it was fixed in
>>> the past year.
>>> Which Solaris release is this?
>>> Thanks,
>>> Cindy
>> Sorry, I see now that you are running s10u7. This CR was fixed in
>> s10u8:
>>
>> http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6462803
>>
>> Cindy
>
> Thank you for that Cindy. I can't seem to find any patch for Solaris
> 10 update 7. I would have thought this significant enough to backport.
> It seems to have hit quite a few people if you Google it.
>
> Dave

Isn't it significant enough to upgrade to Solaris 10 update 8 in order
to be able to install the fix?

Perhaps, if you wait, a patch to update 7 will be forthcoming. Or
perhaps not!
From: ChrisS on 6 Feb 2010 11:15
On Feb 4, 7:14 pm, "Richard B. Gilbert" <rgilber...(a)comcast.net>
wrote:
> David Kirkby wrote:
> > On Feb 4, 5:55 pm, cindy <cindy.swearin...(a)sun.com> wrote:
> >> On Feb 4, 10:48 am, cindy <cindy.swearin...(a)sun.com> wrote:
>
> >>> On Feb 4, 7:46 am, David Kirkby <drkir...(a)gmail.com> wrote:
> >>>> I've hit a problem that a Google search shows others have had before -
> >>>> trying to create a snapshot, one gets a message that the dataset is
> >>>> busy.
> >>>> # zfs snapshot rootpool2/ROOT/S10u7zfsB(a)today
> >>>> cannot create snapshot 'rootpool2/ROOT/S10u7zfsB(a)today': dataset is
> >>>> busy
> >>>> The usual way around this is to unmount then remount the file system..
> >>>> That is not always a practical option though. I can't possibly unmount
> >>>> a file system that many are using. One of the benefits of ZFS is the
> >>>> snapshots. Being told the dataset is busy is not a lot of use.
> >>>> Is there a better solution, or a decent explanation of why this
> >>>> happens?
> >>>> Dave
> >>> Hi Dave,
> >>> One explanation of this problem was a bug, but it was fixed in
> >>> the past year.
> >>> Which Solaris release is this?
> >>> Thanks,
> >>> Cindy
> >> Sorry, I see now that you are running s10u7. This CR was fixed in
> >> s10u8:
>
> >>http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6462803
>
> >> Cindy
>
> > Thank you for that Cindy. I can't seem to find any patch for Solaris
> > 10 update 7. I would have thought this significant enough to backport.
> > It seems to have hit quite a few people if you Google it.
>
> > Dave
>
> Isn't it significant enough to upgrade to Solaris 10 update 8 in order
> to be able to install the fix?
>
> Perhaps, if you wait, a patch to update 7 will be forthcoming.  Or
> perhaps not!

It should be fixed if David applied the Recommended or Sun Alert Patch
Clusters, right? They both contain Kernel patch 1414[4|5]-09. I've
successfully loaded both Alert and Recommended this month, and they
didn't "seem to break anything. (No promises) :-) It should get you
up to ZFS v15, which fixed a similar problem I was having with ZFS
(oh, rebooting after creating a new BE (it left legacy mount entries
in /etc/vfstab), not that you're having this issue, though I was also
using S10u7 at the time. Upate 8 has a bad raidctl command in it (off
of DVD)

First  |  Prev  | 
Pages: 1 2
Prev: Start program from boot, which needs library in non-standard place
Next: Trying to optimize PHP on a T5140