Problem with agregate free space and snapmirroring

[PASCAL_78]

Hello everybody.

I've got 2 FAS 2220 with 8.1.1 7-mode.

I activated asynchronous snapmirroring (on nfs_logs) , and it's working fine, but I can see that agregate free space is going down.

The first time (before snapmirroring), the output was like it:

aggr show_space -h

Aggregate 'aggr0'

    Total space    WAFL reserve    Snap reserve    Usable space       BSR NVLOG           A-SIS          Smtape

           24TB          2479GB             0KB            21TB             0KB             0KB             0KB

Space allocated to volumes in the aggregate

Volume                          Allocated            Used       Guarantee

vol0                                173GB          4258MB          volume

nfs_logs                             21TB           105MB          volume

Aggregate                       Allocated            Used           Avail

Total space                          21TB          4363MB           626GB

Snap reserve                          0KB            0KB             0KB

WAFL reserve                       2479GB            18GB          2460GB

Now :

aggr show_space -h

Aggregate 'aggr0'

    Total space    WAFL reserve    Snap reserve    Usable space       BSR NVLOG           A-SIS          Smtape

           24TB          2479GB             0KB            21TB             0KB             0KB             0KB

Space allocated to volumes in the aggregate

Volume                          Allocated            Used       Guarantee

vol0                                173GB          4653MB          volume

nfs_logs                             21TB          8425MB          volume

Aggregate                       Allocated            Used           Avail

Total space                          21TB            12GB           614GB

Snap reserve                          0KB            11GB             0KB

WAFL reserve                       2479GB            18GB          2460GB

I dont't understand why snap reserve is growing, because agregate scheduling isn't working:

snap sched -A

Aggregate aggr0: 0 0 0

Thanks

[LTDCLSERGAO]

Hello,

Could you give ; df  -Ah and snap list -A

[PASCAL_78]

Hello Said,  this is the output:

df -Ah

Aggregate                total       used      avail capacity

aggr0                     21TB       21TB      614GB      97%

aggr0/.snapshot            0GB       11GB        0GB     ---%

and :

snap list -A

Aggregate aggr0

working...

  %/used       %/total  date          name

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

17% (17%)    0% ( 0%)  Nov 30 09:00  hourly.0

20% ( 4%)    0% ( 0%)  Nov 30 00:00  nightly.0

22% ( 4%)    0% ( 0%)  Nov 29 19:00  hourly.1

25% ( 3%)    0% ( 0%)  Nov 29 14:00  hourly.2

27% ( 4%)    0% ( 0%)  Nov 29 09:00  hourly.3

[LTDCLSERGAO]

strange !!!!

Could you verify the date at which you disabled the aggragat snap sched ?

Is there any script taking this snap ?

If no , delete this snap and observe what is going to take place.

[PASCAL_78]

I disabled agregate snap sched on Nov 30, to see if agregate snapshots will be the same size in the future or not.

But i noticed that it was the same problem and agregate snapshot are using more and more space, so the agregate free space is goig down, whereas no agregate scheduling is active.

[LTDCLSERGAO]

OK.

Snapshots  keep track of the blocks that have changed  between snapshots and the current filesystem. The  changes  introduce to a filesystem can  increases  the snapshot size.

If you dont' need those snap, delete them.

[PASCAL_78]

Thanks Said.

I deleted all existing agregate snapshots, I'll see tomorow if somme agregate snapshots will be generated or not.

I hope "no".

[PASCAL_78 · Accepted solution]

All is ok now, no more agregate snapshots arre generated.

Thanks you Said.

[uptimenow]

Also note that when you delete aggregate snapshots, not all blocks are freed immediately. There's a WAFL scan running to reclaim the freed aggregate blocks ... This may distort any "df -A" or "aggr show_space" figures you look at.