Okay, after creating 16 volumes on 6.1 machines for snapmirror from 5.3 machines, can I say that there really needs to be a "vol create" option to say "-checksums old". It has been a pain to create all these volumes just to wait for checksum reversion. Installing the filer on 5.3 and upgrading after the initial snapmirror isn't an option either - the destination already has other production volumes - besides, thats a kludge.
Beyond that, has anyone else noticed that checksum reversions seem to be "single tasked"? For instance:
Thu Sep 13 15:46:17 PDT [reactor: worker_thread:error]: A snapmirror transfer from vol0 to vol0 has found that the two volumes have differing checksum states. Data ONTAP is aborting the snapmirror transfer and reverting the checksum state for the destination volume vol0. After Data ONTAP displays a message showing the revert is complete, you can retry the snapmirror transfer. You can use the vol status -c command to see the progress of the reversion. Thu Sep 13 15:46:31 PDT [reactor: worker_thread:error]: A snapmirror transfer from vol0 to vol2 has found that the two volumes have differing checksum states. Data ONTAP is aborting the snapmirror transfer and reverting the checksum state for the destination volume vol2. After Data ONTAP displays a message showing the revert is complete, you can retry the snapmirror transfer. You can use the vol status -c command to see the progress of the reversion.
# rsh reactor vol status -c vol0 Volume Checksum style Checksum status vol0 zoned Checksums reverting (98% done) # rsh reactor vol status -c vol2 Volume Checksum style Checksum status vol2 zoned Checksums reverting (0% done)
Just an FYI for other folks - this takes a few hours to complete for each volume when the volumes are ~500GB.
-- Jeff