Hi list, some months ago I spoke with an zfs expert on a Sun Storage event. He told it''s possible to grow a zpool by replacing every single disk with a larger one. After replacing and resilvering all disks of this pool zfs will provide the new size automatically. Now I found time to check that and I was not able to grow the pool. It''s still the same size as before. Maybe I missed one step? Or this functionality is not implemented yet? For your info: I tested this using VMWare and a Solaris 10u6 Any comments or ideas? Thank you in advance... Tobias
Tobias Exner wrote:> Hi list, > > some months ago I spoke with an zfs expert on a Sun Storage event. > > He told it''s possible to grow a zpool by replacing every single disk > with a larger one. > After replacing and resilvering all disks of this pool zfs will > provide the new size automatically. > > > Now I found time to check that and I was not able to grow the pool. > It''s still the same size as before. > > Maybe I missed one step? Or this functionality is not implemented yet?Can you post the steps that you took? Output of ''zpool history'' would be helpful.> > For your info: > > I tested this using VMWare and a Solaris 10u6 > > > > Any comments or ideas?Did you export and then import the pool afterwards? This step is needed from recollection for the system to make the new capacity available in the pool. I did this about 3-4 weeks ago on a mirror and this step was required to make the change in capacity seen by the system. This was on Solaris 10 10/08 which is the same as S10u6.> > > Thank you in advance... > > > Tobias > > > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
On Mon, Aug 03, 2009 at 11:51:59AM +0200, Tobias Exner wrote:> Hi list, > > some months ago I spoke with an zfs expert on a Sun Storage event. > > He told it''s possible to grow a zpool by replacing every single disk > with a larger one. > After replacing and resilvering all disks of this pool zfs will provide > the new size automatically. > > > Now I found time to check that and I was not able to grow the pool. It''s > still the same size as before. > > Maybe I missed one step? Or this functionality is not implemented yet? > > For your info: > > I tested this using VMWare and a Solaris 10u6 > > > > Any comments or ideas?Did you "zpool set autoexpand=on <yourpool>"? It''s off by default... Kurt
*Hi Kurt, Hi Scott, there''s no such property available. There''s no entry in the manpages, too.* /bash-3.00# zpool set autoexpand=on testpool cannot set property for ''testpool'': invalid property ''autoexpand''/ *Maybe a problem of the zfs version? Here''s my "zpool history" * /Verlauf f?r ''testpool'': 2009-08-03.11:14:27 zpool create testpool raidz1 c1t1d0 c1t2d0 c1t3d0 2009-08-03.11:20:48 zpool offline testpool c1t1d0 2009-08-03.11:24:19 zpool replace testpool c1t1d0 c1t4d0 2009-08-03.11:28:02 zpool offline testpool c1t2d0 2009-08-03.11:28:31 zpool replace testpool c1t2d0 c1t5d0 2009-08-03.11:29:07 zpool offline testpool c1t3d0 2009-08-03.11:29:21 zpool replace testpool c1t3d0 c1t6d0 bash-3.00# zfs list testpool NAME USED AVAIL REFER MOUNTPOINT testpool 230M 121M 230M /testpool/ *I replaced 3 virtual disks ( 200mbytes ) with 3 virtual disks ( 400mbytes ) regards, Tobias * Kurt Schreiner schrieb:> On Mon, Aug 03, 2009 at 11:51:59AM +0200, Tobias Exner wrote: > >> Hi list, >> >> some months ago I spoke with an zfs expert on a Sun Storage event. >> >> He told it''s possible to grow a zpool by replacing every single disk >> with a larger one. >> After replacing and resilvering all disks of this pool zfs will provide >> the new size automatically. >> >> >> Now I found time to check that and I was not able to grow the pool. It''s >> still the same size as before. >> >> Maybe I missed one step? Or this functionality is not implemented yet? >> >> For your info: >> >> I tested this using VMWare and a Solaris 10u6 >> >> >> >> Any comments or ideas? >> > Did you "zpool set autoexpand=on <yourpool>"? It''s off by default... > > Kurt > > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090803/af8925c7/attachment.html>
Hi Francois, I just tried it and it''s done. Thank you very much! *my "zpool history"* Verlauf f?r ''testpool'': 2009-08-03.11:14:27 zpool create testpool raidz1 c1t1d0 c1t2d0 c1t3d0 2009-08-03.11:20:48 zpool offline testpool c1t1d0 2009-08-03.11:24:19 zpool replace testpool c1t1d0 c1t4d0 2009-08-03.11:28:02 zpool offline testpool c1t2d0 2009-08-03.11:28:31 zpool replace testpool c1t2d0 c1t5d0 2009-08-03.11:29:07 zpool offline testpool c1t3d0 2009-08-03.11:29:21 zpool replace testpool c1t3d0 c1t6d0 2009-08-03.13:26:00 zpool export testpool 2009-08-03.13:26:15 zpool import testpool *Before export/import* bash-3.00# zfs list testpool NAME USED AVAIL REFER MOUNTPOINT testpool 230M 121M 230M /testpool *After export/import* bash-3.00# zfs list testpool NAME USED AVAIL REFER MOUNTPOINT testpool 230M 529M 230M /testpool regards, Tobias Francois schrieb:> Tobias Exner wrote: >> Hi list, >> >> some months ago I spoke with an zfs expert on a Sun Storage event. >> >> He told it''s possible to grow a zpool by replacing every single disk >> with a larger one. >> After replacing and resilvering all disks of this pool zfs will >> provide the new size automatically. >> >> >> Now I found time to check that and I was not able to grow the pool. >> It''s still the same size as before. >> >> Maybe I missed one step? Or this functionality is not implemented yet? >> >> For your info: >> >> I tested this using VMWare and a Solaris 10u6 >> >> >> >> Any comments or ideas? > > Did you tried to export/import the pool ? > > >> >> >> Thank you in advance... >> >> >> Tobias >> >> >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss >> >> > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090803/317c9097/attachment.html>
Hi Tobias, On Mon, Aug 03, 2009 at 01:24:35PM +0200, Tobias Exner wrote:> > there''s no such property available. > There''s no entry in the manpages, too. > > > bash-3.00# zpool set autoexpand=on testpool > cannot set property for ''testpool'': invalid property ''autoexpand'' > > Maybe a problem of the zfs version?Hm, yes, I think so. According to the zfsadminguide from June 2009, page 16, this is a new feature in build 117. On earlier builds an export/import seems to be needed. (Not tried this as the machine I replaced the old 36G disks on runs build 117 w/ zfs pool version 15.) Kurt
Ahh, I understand. The autoexpand is a feature which allows to grow the filesystem on the fly. That''s great but export/import is good enough for the moment. thanks, Tobias Kurt Schreiner schrieb:> Hi Tobias, > > On Mon, Aug 03, 2009 at 01:24:35PM +0200, Tobias Exner wrote: > >> there''s no such property available. >> There''s no entry in the manpages, too. >> >> >> bash-3.00# zpool set autoexpand=on testpool >> cannot set property for ''testpool'': invalid property ''autoexpand'' >> >> Maybe a problem of the zfs version? >> > Hm, yes, I think so. According to the zfsadminguide from June 2009, page 16, > this is a new feature in build 117. On earlier builds an export/import > seems to be needed. (Not tried this as the machine I replaced the old 36G > disks on runs build 117 w/ zfs pool version 15.) > > Kurt > > >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20090803/d4d9735e/attachment.html>
Kurt Schreiner wrote: Hi Tobias, On Mon, Aug 03, 2009 at 01:24:35PM +0200, Tobias Exner wrote: there''s no such property available. There''s no entry in the manpages, too. bash-3.00# zpool set autoexpand=on testpool cannot set property for ''testpool'': invalid property ''autoexpand'' Maybe a problem of the zfs version? this is a new feature in build 117. On earlier builds an export/import seems to be needed. (Not tried this as the machine I replaced the old 36G disks on runs build 117 w/ zfs pool version 15.) It also happens (at least in snv_113) just by rebooting (which can be really bad news if you need it not to happen). -- Andrew _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss