SnapMirror Transfer Pauses Randomly

[nicholsongc]

I have four relationships between the same two ONTAP 8.3.2 clusters.  Three relationships update every 4 hours in an acceptible time frame.

One of the four starts transferring then pauses....  sometimes for 20-30 minutes... then continues... then pauses some more....

Over the course of the day the average time to complete the SLOW relationship is over 1 hour.

Average for the other transfers are all under 23 minutes.

Latest transfer stats:

Source:volA --> Dest:volA_mirror average bytes transferred is 743MB and takes an average of 26 seconds

Source:volB --> Dest:volB_mirror average bytes transferred is 22.6GB and takes an average of 22 minutes 6 seconds

Source:volC --> Dest:volC_mirror average bytes transferred is 5GB and takes an average of 1 hour and 9 minutes

Source:volD --> Dest:volD_mirror average bytes transferred is 12.6GB and takes an average of 4 minutes 30 seconds

All fours source volumes are on the same node/aggregate, all four destination volumes are on the same node/aggregate.

What could be causing 1 of the 4 to drag on for so long... some WAFL process, etc.??

[GidonMarcus]

i'm not aware for a knwon reason. and can't see a fixed one in

https://mysupport.netapp.com/NOW/cgi-bin/relcmp.on?notfirst=Go%21&rrel=8.3.2&rels=8.3.2P12&what=fix

you might find it in the logs:

EMS

/etc/log/snapmirror_audit*
/etc/log/snapmirror_error*
/etc/log/mlog/snapmirror*

but it's a pain to get them....  (https://kb.netapp.com/app/answers/answer_view/a_id/1001462, https://kb.netapp.com/app/answers/answer_view/a_id/1033620)

you better just trigger an ASUP to NetApp, look for it in ActiveIQ (My Autosupport) or open a support case and let them search.

if nothing found there. you would likely need to capture a perfstat on both ends while it happening. and a long packettrace on the dest (or both) with only the headers (https://kb.netapp.com/app/answers/answer_view/a_id/1032808/loc/en_US)