Prev: ZFS - Backup configuration
Next: How to use a dvd image to install Solaris 10 on E 250/E 450
From: Zfs.. on 18 Jan 2010 17:22
On Jan 18, 10:11 pm, Darren Dunham <darren.dun...(a)gmail.com> wrote:
> On Jan 18, 11:32 am, "Zfs.." <cian.scrip...(a)gmail.com> wrote:
>
> > Thanks for the reply Darren. I do understand that I'm issuing a force
> > of the pool, and of course you must be sure that if you use this flag
> > you better know what you are doing. However, with zfs, this is the
> > ONLY way to import the pool to another host which seems funny to me.
> > Correct me if I'm wrong, but this has been my experience with zfs.
>
> Not if it has been exported.
>
> While the pool is in use (imported), it marks the pool as in use by
> the particular host.  Any other host trying to import will fail. If
> the original host exports the pool that lock is cleared.
>
> There's really not any good way to tell the difference between
> "really" in use, and "was in use a while ago, but not in use any
> longer".  The system simply sets and clears the in-use flag.
>
> > If you issue zpool import mypool on a pool that was accessed on
> > another system it will tell you to use the -f option.
>
> If it hasn't been explicitly exported, yes.  If you're expecting to
> use this on multiple systems, doing a clean export should be part of
> your workflow so that you don't have to force the import.
>
> > This is something I think Sun need to look at, surely there has to be
> > another way to do this without issuing -f to the import.
>
> Assuming you're exporting it, there should be no problem.
>
> --
> Darren

Darren,

I'm suffering from amnesia.. Yes, yes, yes, This is what we do in our
setup. ( Doh ! )

When we where looking at all the pools today on the system from the
second node they suggested that they can be imported via -f option. I
think we where second guessing ourselves somewhat.

When we fail over our pools, we try first zpool import, then zpool
import -f, I might add another check in to see if the pool is actually
imported on the second node, maybe via a remote ssh command.

Thanks a million Darren, and all the other contributors to this
thread, you've helped me kick start my brain somewhat.

I do still think that a lockhost property could be useful in zfs, but
hey we can't have everything I suppose !
From: Chris Ridd on 19 Jan 2010 01:48
On 2010-01-18 22:22:50 +0000, Zfs.. said:

> When we fail over our pools, we try first zpool import, then zpool
> import -f, I might add another check in to see if the pool is actually
> imported on the second node, maybe via a remote ssh command.
>
> Thanks a million Darren, and all the other contributors to this
> thread, you've helped me kick start my brain somewhat.
>
> I do still think that a lockhost property could be useful in zfs, but
> hey we can't have everything I suppose !

I don't think that would add anything. Your pool failover logic would
still need to have a "is this *really* being used" step, and now you
need to worry about "is this *really* locked AND is it *really*
imported somewhere else".
--
Chris

From: hume.spamfilter on 19 Jan 2010 12:22
Zfs.. <cian.scripter(a)gmail.com> wrote:
> I do still think that a lockhost property could be useful in zfs, but
> hey we can't have everything I suppose !

That strikes me as a potential disaster if the "locked" host died and your
administrators are desperately trying to bring the pool back up on another
host to recover the data.

Then you'd need a "-f -f" flag, for "yes, I really mean force the mount".

But then somebody would "import -f -f" the pool onto two hosts and be
surprised when it was corrupted, and they'd make a post on comp.unix.solaris
saying that perhaps there should be a "lockhost_yeahreally" property to
keep them from doing that...

--
Brandon Hume - hume -> BOFH.Ca, http://WWW.BOFH.Ca/
From: Zfs.. on 19 Jan 2010 14:40
On Jan 19, 5:22 pm, hume.spamfil...(a)bofh.ca wrote:
> Zfs.. <cian.scrip...(a)gmail.com> wrote:
> > I do still think that a lockhost property could be useful in zfs, but
> > hey we can't have everything I suppose !
>
> That strikes me as a potential disaster if the "locked" host died and your
> administrators are desperately trying to bring the pool back up on another
> host to recover the data.
>
> Then you'd need a "-f -f" flag, for "yes, I really mean force the mount".
>
> But then somebody would "import -f -f" the pool onto two hosts and be
> surprised when it was corrupted, and they'd make a post on comp.unix.solaris
> saying that perhaps there should be a "lockhost_yeahreally" property to
> keep them from doing that...
>
> --
> Brandon Hume    - hume -> BOFH.Ca,http://WWW.BOFH.Ca/

Ok I take your point.. :-)
First  |  Prev  | 
Pages: 1 2 3 4
Prev: ZFS - Backup configuration
Next: How to use a dvd image to install Solaris 10 on E 250/E 450