ONTAP Discussions

SVM_DR Snapshots not deleted on source



I've a question regarding the Snapshot (sm_created) used by SnapMirror for replication. I've probably misunderstood somehting but I cannot figure out what.


I have a cascaded replication: Cluster A  >  Cluster B (SVM_DR)  >  Cluster C (SnapVault - long term of Weekly)


The source volume (on Cluster A) has a snapshot policy set to "default", unchanged:


Vserver: cluster1
Number of IsPolicy Name Schedules Enabled Comment
------------------------ --------- ------- ----------------------------------
default 4 true Default policy with hourly, daily & weekly schedules.
Schedule Count Prefix SnapMirror Label
---------------------- ----- ---------------------- -------------------
hourly     6      hourly        -
daily      2      daily         daily
weekly     2      weekly        weekly
5min      12      5min          5min


So no "sm_created" label.  At this time, the SVM_DR replication ran for a couple of days and on my source volume (cluster 1) I still have the following snapshots:


snap list -vserver volume snapshot snapmirror-label
------------------------------------------------------------------ ----------------
vs1 cascade_src1 vserverdr.0.d3c6b107-390d-11e9-a1d2-005056b013db.2019-02-25_190000 sm_created

[...] Ouput truncated

vs1 cascade_src1 vserverdr.2.d3c6b107-390d-11e9-a1d2-005056b013db.2019-02-27_122500 sm_created
vs1 cascade_src1 vserverdr.0.d3c6b107-390d-11e9-a1d2-005056b013db.2019-02-27_123000 sm_created
12 entries were displayed.


12 Snapshots are kept, the retention count for 5min snapshot is 12 and the Snapmirror schedule for the SVM_DR is indeed 5min. 

If the retention count is defined by the snamirror label, why is is this using 5min instead of "sm_created" which is the real label ?


Do you have any best practices or recommandation to use in this kind of situation ?


Thanks for your help !




@brasseurm  Let me know if you are still looking for help, I will help you find an expert who can answer to your query.