ONTAP Discussions

Aggr snapshot autodelete

rajdeepsengupta
4,953 Views

I have an aggregate in a 3240 system, which is highly loaded, I mean many volumes in that aggr shows above 50msec read latency and above 10msec write latency. But I also find that at times a autodelete snapshot gets created in that aggr, which overshoots the aggr snapshot.

Two questions:

1. Whay this autodelete aggr snapshot are created, I mean under what conditon.

2. Since I have performance issue anyway, so should I disable the snapshot creation of that aggr?

3. Is there any valid reasons to keep aggr snapshot reserve, I mean I do not use syncmirror.

A little detail explanation will be appreciated.

Thanks

Rajdeep

8 REPLIES 8

aborzenkov
4,953 Views

Could you explain what do you mean under “autodelete snapshot gets created”? What is the difference between “autodelete snapshot” and normal snapshot?

rajdeepsengupta
4,953 Views

Normal snapshot is what you can create directly or thru snashot schedule, like hourly.0 daily.1 etc. But autodelete snapshot is created by the system. Now at what condition the system creates such snapshot is not know.

If you do snap list -A aggr, you can see this snapshot, and the system will delete it also after a certain period automatically.

aborzenkov
4,953 Views

These snapshots are also created by schedule. See “snap sched -A”. They can be used to recover data that has been deleted (like destroyed flexible volume), although that probably requires NetApp support. You can disable schedule if you think you won’t need them.

rajdeepsengupta
4,953 Views

No, those snapshots which are created thru schedules are named as horly.0, nightly.0 or weekly.0, they are not created as "autodelete".

My snapsched for aggr is like..

Aggregate aggr1: 0 0 6@1,4,8,12,16,20

aborzenkov
4,953 Views

It would be easier if you show “snap list -A” and tell, which snapshot do you mean.

rajdeepsengupta
4,953 Views

Output of snap list -A

Aggregate aggr1

working......

  %/used       %/total  date          name

----------  ----------  ------------  --------

  3% ( 3%)    2% ( 2%)  Aug 30 13:05  autodelete_base

rajdeepsengupta
4,953 Views

Yes, this article is helpful.

Just wanted to know after creation of the autodelete snapshot, what it actually does, I mean the internal logic. I was finding the snapshot slowly reserving the space from 0% to 3%. So internally it must be saving the data which volumes deleted.

Also during this phase the volume read/write latecny was too high.

After I removed the snapshot and removed the space reserve, the latency improved. Not 100% sure becuase of this, but it seems.

Public