Hello zfs-discuss, What happens after a reboot to l2arc? I mean after pool is imported will zfs try to still check for data in a l2arc or will it assume it is empty and will start to warm it up? Since data are checksummed on l2arc one could argue you it should be possible to still check with l2arc? it could be especially useful in cluster config - you switch your database with your pool to another node and l2arc is already warm (hot!). ? -- Best regards, Robert Milkowski mailto:milek at task.gda.pl http://milek.blogspot.com
Robert Milkowski wrote:> Hello zfs-discuss, > > What happens after a reboot to l2arc? I mean after pool is imported > will zfs try to still check for data in a l2arc or will it assume it > is empty and will start to warm it up?Empty and warm it up.> Since data are checksummed on l2arc one could argue you it should be > possible to still check with l2arc?The checksums are in memory not on disk.> it could be especially useful in cluster config - you switch your > database with your pool to another node and l2arc is already warm > (hot!).Indeed but that would require putting the checksums on disk. I have heard rumour of a plan to do that, not sure if your example is the motivation or not. -- Darren J Moffat
Darren J Moffat wrote:> Robert Milkowski wrote: > >> Hello zfs-discuss, >> >> What happens after a reboot to l2arc? I mean after pool is imported >> will zfs try to still check for data in a l2arc or will it assume it >> is empty and will start to warm it up? >> > > Empty and warm it up. > > >> Since data are checksummed on l2arc one could argue you it should be >> possible to still check with l2arc? >> > > The checksums are in memory not on disk. > > >> it could be especially useful in cluster config - you switch your >> database with your pool to another node and l2arc is already warm >> (hot!). >> > > Indeed but that would require putting the checksums on disk. > > I have heard rumour of a plan to do that, not sure if your example is > the motivation or not. > >See CR6662467 http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6662467 -- richard