ONTAP Discussions

Deswizzle - how to read status

FrostByte
12,733 Views

I've read through the old deswizzle article and looked around online for information to no avail.

We have a 12TB volume that we replicate to DR with a change rate of around 40 GB per 4 hour replication.  Recently, there was an abnormal load on the filer an that may have prevented deswizzle from completing.  We are getting a finger pointed at us for slow performance.  My question is how do I read the output of the following command:

wafl scan status

Scan id                   Type of scan     progress

    4573             volume deswizzling     snap 28, inode 30072 of 32781. level 1 of normal files. Totals: Normal files: L1:234113/1203654 L2:0/1316911 L3:0/1319542 L4:0/8       Inode file: L0:0/0 L1:0/0 L2:0/0 L3:0/0 L4:0/0

I'm assuming that 'snap 28' coorisponds to the snapshot:  filer(0118073998)_asmv_naf1b1_edb_epicprdvg.28 (busy,snapmirror,vclone)

It appears that the inode 30072 of 32781 might be an indication of where we are in the process as the first number is growing whiile the second is not.

Any help would be appreciated.

1 ACCEPTED SOLUTION

pascalduk
12,733 Views

You can find the snapshot name belonging to the "snapid" by running the following command in advance mode:

snap status <volume>

The filename belonging to an inode number can be found using e.g. an nfs mount to volume and the find command.

View solution in original post

6 REPLIES 6

shaunjurr
12,733 Views

Perhaps you could include some information as to the filer model and ONTap version you are using.

Performance problems probably should include some sort of sysstat or perfstat information or it is going to be hard help.

FrostByte
12,733 Views

Sorry, we are running 8.0.1p2 going from a 6080 to a 6030.  Sysstat looks good and everything (now) appears to be healthy and I suspect that the deswizzle will finish relatively soon.  At this point in time, my goal was just to understand the output of the command.

pascalduk
12,734 Views

You can find the snapshot name belonging to the "snapid" by running the following command in advance mode:

snap status <volume>

The filename belonging to an inode number can be found using e.g. an nfs mount to volume and the find command.

thomas_glodde
12,733 Views

Hi,

deswizzling will map the physical blocks for a snapmirror replication. the filesystem on filer1 maps to physical blocks. upon replication to filer2, it needs to remap the filesystem to the "new" physical blocks on filerb as disks and their layouts etc. differ. this process can, espacially after snapmirror initialize or after a huge delta replication, take plenty of time and will generate load on the system. you need to unterstand that deswizzle is run everytime after a sucessfull snapmirror repliaction and it will start all over again if you delete a snapshot or if your nexdt replication takes place before finishing the scan. this might force you to stop replication for a amount of time the deswizzle needs to properly finish and then continue with snapmirror repliaction since once it completed, it will only need to process the delta data.

please get in touch with netapp global support if you require further assistance on that matter or if your performance is still impacted.

To actualy answer your question, yes, the first number is rising and deswizzle will finish if the first number reaches the second number 😉 after that your load, especialy your disk utilization, will drop.

Kind regards,

Thomas

FrostByte
12,733 Views

Thanks to all. It is working on the 'oldest' snapshot excluding the two that are a month old that were taken for flexclone purposes. Now, I can keep an eye on it and know if we are making progress towards being caught up. There was an unruly process hogging all disk IO and that likely caused us to get behind the curve.

I did call in a ticket yesterday. I had to spell deswizzle to the level one support person who then proceeded to read me a TR note on it, waste an hour of my time reading documents, and then agree to call back with an update by end of business (5 hours) later. Yea, I didn't hear anything from him.

Fortunately, we have a good community and I appreciate the help

ijm2024
12,183 Views

We have issue where AGGR block reclamation has not yet run on SNAPMIRROR destination filer, since it has not returned space back from deleted snapshots on PRIMARY volume.

Noted that the MIRROR VOLS have been actively deswizzling on 3 day old snapshots, where AGGR w/MIRRORED VOLS is 99% full and PRIMARY AGGR is only 89% full. What support team did you get to help with the deswizzle , block reclamations?

Public