Hi Tomas
I understand from previous post http://www.mail-archive.com/zfs-discuss@opensolaris.org/msg36914.html that if the data gets invalidated, the l2arc size that is shown by zpool iostat is the one that changed (always growing because of COW) not the actual size shown by kstat which represent the size of the up to date data in l2arc. My only conclusion here to this fluctuation in kstat l2_size is the fact that data has indeed invalidated and did not made it back to l2arc from the tail of ARC !!! Am I right ???? On Fri, Apr 9, 2010 at 4:33 PM, Tomas Ögren <st...@acc.umu.se> wrote: > On 09 April, 2010 - Abdullah Al-Dahlawi sent me these 27K bytes: > > > Hi all > > > > I ran an OLTP-Filebench workload > > > > I set Arc max size = 2 gb > > l2arc ssd device size = 32gb > > workingset(dataset) = 10gb , 10 files , 1gb each > > > > after running the workload for 6 hours and monitoring kstat , I have > noticed > > that l2_size from kstat has reached 10gb which is great . however, > l2_size > > started to drop all the way to 7gb !!!! which means that the workload > will > > go back to the HDD to retirive some data that are no longer on l2arc > device > > . > > > > I understand that l2arc size reflected by zpool iostat is much larger > > becuase of COW and l2_size from kstat is the actual size of l2arc data. > > > > so can any one tell me why I am loosing my workingset from l2_size actual > > data !!! > > Maybe the data in the l2arc was invalidated, because the original data > was rewritten? > > /Tomas > -- > Tomas Ögren, st...@acc.umu.se, > http://www.acc.umu.se/~stric/<http://www.acc.umu.se/%7Estric/> > |- Student at Computing Science, University of Umeå > `- Sysadmin at {cs,acc}.umu.se > -- Abdullah Al-Dahlawi PhD Candidate George Washington University Department. Of Electrical & Computer Engineering ---- Check The Fastest 500 Super Computers Worldwide http://www.top500.org/list/2009/11/100
_______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss