Hi, This is what we are trying to understand. Luns are presented to the CDOM and then we create the zpool on them. On top of the zpool zvol is created and then it is presented to the GDOM. zpool list | grep testpool testpool 12G 114K 12.0G 0% ONLINE - zfs list | grep testpool testpool 110K 11.8G 18K /testpool *zfs create -V 10g testpool/testvol* zfs list | grep testpool testpool 10.0G 1.81G 18K /testpool testpool/testvol 10G 11.8G 16K - In the gdom we create the UFS file system using newfs on this disk and then we copy 4GB of data on that. */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* The disk space used increases. zpool list | grep testpool zfs list | grep testpool zpool list | grep testpool testpool 12G 4.49G 7.51G 37% ONLINE - zfs list | grep testpool testpool 10.0G 1.81G 18K /testpool testpool/testvol 10G 7.32G 4.49G - */dev/dsk/c0d1s0 9.8G 4.3G 5.5G 44% /test* After this we delete all the things in the /test directory, so that the disk space is freed. */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* This doesn?t gets reflected on the zfs list and it still shows the same size. zpool list | grep testpool testpool 12G 4.49G 7.51G 37% ONLINE - zfs list | grep testpool testpool 10.0G 1.81G 18K /testpool testpool/testvol 10G 7.32G 4.49G - Thanks for the help in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20100225/fd1aaac6/attachment.html>
Hi see my response in line. On Thu, Feb 25, 2010 at 2:05 AM, Milan Shah <milanmukulshah at gmail.com>wrote:> Hi, > > This is what we are trying to understand. > > Luns are presented to the CDOM and then we create the zpool on them. > > On top of the zpool zvol is created and then it is presented to the GDOM. > > zpool list | grep testpool > > testpool 12G 114K 12.0G 0% ONLINE - > > > > zfs list | grep testpool > > testpool 110K 11.8G 18K /testpool > > > > *zfs create -V 10g testpool/testvol* > > > > zfs list | grep testpool > > testpool 10.0G 1.81G 18K /testpool > > testpool/testvol 10G 11.8G 16K - > > > > In the gdom we create the UFS file system using newfs on this disk and then > we copy 4GB of data on that. > > */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* > > The disk space used increases. > > zpool list | grep testpool > > zfs list | grep testpool > > > > zpool list | grep testpool > > testpool 12G 4.49G 7.51G 37% ONLINE - > > > > zfs list | grep testpool > > testpool 10.0G 1.81G 18K /testpool > > testpool/testvol 10G 7.32G 4.49G - > > > > */dev/dsk/c0d1s0 9.8G 4.3G 5.5G 44% /test* > > > > After this we delete all the things in the /test directory, so that the > disk space is freed. > > */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >the act of deleting files in UFS simply does a few accounting changes to the filesystem thus has no affect on the blocks in ZFS volume, and in some cases could actually make the zvol space grow. The only possible way to have ZFS storage shrink would be to have compression enabled on the zvol, and then over write the files with zero''s and even this wouldn''t necessarily shrink the volume due to ZFS''s copy on write storage. James Dickens http://uadmin.blogspot.com> This doesn?t gets reflected on the zfs list and it still shows the same > size. > > zpool list | grep testpool > > testpool 12G 4.49G 7.51G 37% ONLINE - > > > > zfs list | grep testpool > > testpool 10.0G 1.81G 18K /testpool > > testpool/testvol 10G 7.32G 4.49G - > > > > Thanks for the help in advance. > > _______________________________________________ > 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/20100225/15da89db/attachment.html>
Hi james, thanks for the reply, stating this is there a way i can restirct the size of the zvol. So if i have the zvol of 10 GB on the CDOM, which is presented to the LDOM as the disk and then we create the UFS file system on that, but this grows with time and we eve see the situation where it grows more then that of the actual zvol size and uses the free size from the zpool. We want to restrict the growth of the zvol to a specific size which is much lesser then that of the zpool. Thanks Milan On Thu, Feb 25, 2010 at 12:17 PM, James Dickens <jamesd.wi at gmail.com> wrote:> Hi > > see my response in line. > > On Thu, Feb 25, 2010 at 2:05 AM, Milan Shah <milanmukulshah at gmail.com>wrote: > >> Hi, >> >> This is what we are trying to understand. >> >> Luns are presented to the CDOM and then we create the zpool on them. >> >> On top of the zpool zvol is created and then it is presented to the GDOM. >> >> zpool list | grep testpool >> >> testpool 12G 114K 12.0G 0% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 110K 11.8G 18K /testpool >> >> >> >> *zfs create -V 10g testpool/testvol* >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 11.8G 16K - >> >> >> >> In the gdom we create the UFS file system using newfs on this disk and >> then we copy 4GB of data on that. >> >> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >> >> The disk space used increases. >> >> zpool list | grep testpool >> >> zfs list | grep testpool >> >> >> >> zpool list | grep testpool >> >> testpool 12G 4.49G 7.51G 37% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 7.32G 4.49G - >> >> >> >> */dev/dsk/c0d1s0 9.8G 4.3G 5.5G 44% /test* >> >> >> >> After this we delete all the things in the /test directory, so that the >> disk space is freed. >> >> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >> > > the act of deleting files in UFS simply does a few accounting changes to > the filesystem thus has no affect on the blocks in ZFS volume, and in some > cases could actually make the zvol space grow. The only possible way to have > ZFS storage shrink would be to have compression enabled on the zvol, and > then over write the files with zero''s and even this > wouldn''t necessarily shrink the volume due to ZFS''s copy on write storage. > > > James Dickens > http://uadmin.blogspot.com > > > > >> This doesn?t gets reflected on the zfs list and it still shows the same >> size. >> >> zpool list | grep testpool >> >> testpool 12G 4.49G 7.51G 37% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 7.32G 4.49G - >> >> >> >> Thanks for the help in advance. >> >> _______________________________________________ >> 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/20100225/660883f3/attachment.html>
Hi james, Thanks for the reply, stating this is there a way i can restrict the size of the zvol. So if i have the zvol of 10 GB on the CDOM, which is presented to the LDOM as the disk and then we create the UFS file system on that, but this grows with time and we eve see the situation where it grows more then that of the actual zvol size and uses the free size from the zpool. We want to restrict the growth of the zvol to a specific size which is much lesser then that of the zpool. Thanks Milan On Thu, Feb 25, 2010 at 12:17 PM, James Dickens <jamesd.wi at gmail.com> wrote:> Hi > > see my response in line. > > On Thu, Feb 25, 2010 at 2:05 AM, Milan Shah <milanmukulshah at gmail.com>wrote: > >> Hi, >> >> This is what we are trying to understand. >> >> Luns are presented to the CDOM and then we create the zpool on them. >> >> On top of the zpool zvol is created and then it is presented to the GDOM. >> >> zpool list | grep testpool >> >> testpool 12G 114K 12.0G 0% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 110K 11.8G 18K /testpool >> >> >> >> *zfs create -V 10g testpool/testvol* >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 11.8G 16K - >> >> >> >> In the gdom we create the UFS file system using newfs on this disk and >> then we copy 4GB of data on that. >> >> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >> >> The disk space used increases. >> >> zpool list | grep testpool >> >> zfs list | grep testpool >> >> >> >> zpool list | grep testpool >> >> testpool 12G 4.49G 7.51G 37% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 7.32G 4.49G - >> >> >> >> */dev/dsk/c0d1s0 9.8G 4.3G 5.5G 44% /test* >> >> >> >> After this we delete all the things in the /test directory, so that the >> disk space is freed. >> >> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >> > > the act of deleting files in UFS simply does a few accounting changes to > the filesystem thus has no affect on the blocks in ZFS volume, and in some > cases could actually make the zvol space grow. The only possible way to have > ZFS storage shrink would be to have compression enabled on the zvol, and > then over write the files with zero''s and even this > wouldn''t necessarily shrink the volume due to ZFS''s copy on write storage. > > > James Dickens > http://uadmin.blogspot.com > > > > >> This doesn?t gets reflected on the zfs list and it still shows the same >> size. >> >> zpool list | grep testpool >> >> testpool 12G 4.49G 7.51G 37% ONLINE - >> >> >> >> zfs list | grep testpool >> >> testpool 10.0G 1.81G 18K /testpool >> >> testpool/testvol 10G 7.32G 4.49G - >> >> >> >> Thanks for the help in advance. >> >> _______________________________________________ >> 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/20100225/41e3f949/attachment.html>
On Thu, Feb 25, 2010 at 11:28 AM, Milan Shah <milanmukulshah at gmail.com>wrote:> Hi james, > > Thanks for the reply, stating this is there a way i can restrict the size > of the zvol."Quotas cannot be set on volumes, as the volsize property acts as an implicit quota." from the zfs manpage You could try reseting volume size property but this will probably have a negative impact on what ever fileysystem you have stored on the volume. James Dickens http://uadmin.blogspot.com> So if i have the zvol of 10 GB on the CDOM, which is presented to the LDOM > as the disk and then we create the UFS file system on that, but this grows > with time and we eve see the situation where it grows more then that of the > actual zvol size and uses the free size from the zpool. > > We want to restrict the growth of the zvol to a specific size which is much > lesser then that of the zpool. > > Thanks > Milan > > On Thu, Feb 25, 2010 at 12:17 PM, James Dickens <jamesd.wi at gmail.com>wrote: > >> Hi >> >> see my response in line. >> >> On Thu, Feb 25, 2010 at 2:05 AM, Milan Shah <milanmukulshah at gmail.com>wrote: >> >>> Hi, >>> >>> This is what we are trying to understand. >>> >>> Luns are presented to the CDOM and then we create the zpool on them. >>> >>> On top of the zpool zvol is created and then it is presented to the GDOM. >>> >>> zpool list | grep testpool >>> >>> testpool 12G 114K 12.0G 0% ONLINE - >>> >>> >>> >>> zfs list | grep testpool >>> >>> testpool 110K 11.8G 18K /testpool >>> >>> >>> >>> *zfs create -V 10g testpool/testvol* >>> >>> >>> >>> zfs list | grep testpool >>> >>> testpool 10.0G 1.81G 18K /testpool >>> >>> testpool/testvol 10G 11.8G 16K - >>> >>> >>> >>> In the gdom we create the UFS file system using newfs on this disk and >>> then we copy 4GB of data on that. >>> >>> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >>> >>> The disk space used increases. >>> >>> zpool list | grep testpool >>> >>> zfs list | grep testpool >>> >>> >>> >>> zpool list | grep testpool >>> >>> testpool 12G 4.49G 7.51G 37% ONLINE - >>> >>> >>> >>> zfs list | grep testpool >>> >>> testpool 10.0G 1.81G 18K /testpool >>> >>> testpool/testvol 10G 7.32G 4.49G - >>> >>> >>> >>> */dev/dsk/c0d1s0 9.8G 4.3G 5.5G 44% /test* >>> >>> >>> >>> After this we delete all the things in the /test directory, so that the >>> disk space is freed. >>> >>> */dev/dsk/c0d1s0 9.8G 10M 9.7G 1% /test* >>> >> >> the act of deleting files in UFS simply does a few accounting changes to >> the filesystem thus has no affect on the blocks in ZFS volume, and in some >> cases could actually make the zvol space grow. The only possible way to have >> ZFS storage shrink would be to have compression enabled on the zvol, and >> then over write the files with zero''s and even this >> wouldn''t necessarily shrink the volume due to ZFS''s copy on write storage. >> >> >> James Dickens >> http://uadmin.blogspot.com >> >> >> >> >>> This doesn?t gets reflected on the zfs list and it still shows the same >>> size. >>> >>> zpool list | grep testpool >>> >>> testpool 12G 4.49G 7.51G 37% ONLINE - >>> >>> >>> >>> zfs list | grep testpool >>> >>> testpool 10.0G 1.81G 18K /testpool >>> >>> testpool/testvol 10G 7.32G 4.49G - >>> >>> >>> >>> Thanks for the help in advance. >>> >>> _______________________________________________ >>> 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/20100225/587e41e6/attachment.html>