On Mon, May 31, 2010 at 01:33:38PM -0400, Page, Jeremy wrote:
Have you tried a reallocate measure? What does a statit look like during the xfer?
The source volume: Tue Jun 1 13:57:56 CEST [filer-hm6e: wafl.reallocate.check.value:info]: Allocation measurement check on '/vol/test1' is 1.
The destination vol after breaking: Tue Jun 1 14:48:00 CEST [filer-hm6e: wafl.reallocate.check.value:info]: Allocation measurement check on '/vol/testcopy' is 1.
Statit shows all cpus are > 90 % idle. Busyest disk is 3 % busy with 3 IO/s. Mostly writing full stripes: 121.65 18 blocks per stripe size 18 19.85 1 blocks per stripe size 1 and the rest <2
Did it ever run properly?
Well. Yes and no.
This used to be a stand alone filer for several years. It must have been a fully functional snapmirror destination at one time but I don't remember. I am sure that during those years it was a fast snapmirror source and NFS server 24*7. Then we moved the head and disks to another location and converted two of these to a fabric metro cluster. It never worked after the conversion. But its partner is.
But someone from NetApp just mailed about options replication.throttle.* and that was the problem.
Was: replication.throttle.enable on replication.throttle.incoming.max_kbs 10000
And now with: replication.throttle.enable off replication.throttle.incoming.max_kbs 125000 problem is over.
We did not know about that option, surely did not set it and must have read over it several times.
Thanks anyway,