I have a question about the snapmirror checkpoint in cDOT831. And I got the answer that the checkpoint is set every 30s from DLs.
But in my test, I cannot see the checkpoint during snapmirroring in some scenario. So not sure about the snapmirror restart checkpoint mechanism.
Here are some logs,
1. During all the transferring for about 15min, the checkpoint is always 0Bytes. The files size is about 10Mbytes or less.
cluster2::*> snapmirror show -fields snapshot-checkpoint,total-progress,status,current-transfer-type,current-throttle -destination-volume snapmirror_dst1
source-path destination-path current-throttle status total-progress snapshot-checkpoint current-transfer-type
------------------------ ------------------------ ---------------- ------------ -------------- ------------------- ---------------------
snap_src:snapmirror_src1 svm_dst1:snapmirror_dst1 819 Transferring 200.1MB 0B update -> during the snapmirror transferring, the checkpoint is always 0Byte
2. Then I tried the big files of 1Gbyte(fsutil file createnew a.txt 1024000000), the checkpoint did update about every 30s.
cluster2::*> date
Node Date Time zone
--------- ------------------------ -------------------------
cluster2-01
Mon Feb 01 08:00:32 2016 Etc/UTC
cluster2::*> snapmirror show -fields snapshot-checkpoint,total-progress,status,current-transfer-type,current-throttle -destination-volume snapmirror_dst1
source-path destination-path current-throttle status total-progress snapshot-checkpoint current-transfer-type
------------------------ ------------------------ ---------------- ------------ -------------- ------------------- ---------------------
snap_src:snapmirror_src1 svm_dst1:snapmirror_dst1 819 Transferring 96.25MB 44.12MB update
cluster2::*> date
Node Date Time zone
--------- ------------------------ -------------------------
cluster2-01
Mon Feb 01 08:01:00 2016 Etc/UTC -> after about 30s
cluster2::*> snapmirror show -fields snapshot-checkpoint,total-progress,status,current-transfer-type,current-throttle -destination-volume snapmirror_dst1
source-path destination-path current-throttle status total-progress snapshot-checkpoint current-transfer-type
------------------------ ------------------------ ---------------- ------------ -------------- ------------------- ---------------------
snap_src:snapmirror_src1 svm_dst1:snapmirror_dst1 819 Transferring 121.1MB 58.62MB update
cluster2::*>
Oppositely, remembers in 7mode, snapmirror will set a restart checkpoint every 5 minutes when the previous inode has completed. That means for small file, you will see the checkpoint every 5min. But for big file, you’ll see the checkpoint till the complete file has done.
So, not sure when will checkpoint be set in cDOT, and whether it is related to the file size or something else.
