ONTAP Discussions

SME logs snapmirrored to DR

BOOSTMR2VA
2,283 Views

Hi guys, i am fairly new to these forums.  Novice netapp user here.

I am re-evaluating our Netapp DR setup after some recent issues with snapmirror, all having to do with limited WAN performance. 

I will have a snapmirror snapshot for a volume, as well as a SME snapshot on the volume.  About 60GB of logs have been adjusted on the volume (technically the LUN on the volume)

Snapmirror snapshot 6/12/12 - 2 GB in size

SME snapshot 6/12/12 - 60 GB in size

If I remove the SME snapshot, will the snapmirror snapshot still be 62GB cumulative size?  I would imagine so because there were still 60GB changes in the log files on the volume.  I assume Netapp is being smart in not copying 60GB x 2 to the snapmirror destination.

2 REPLIES 2

martin_fisher
2,283 Views

Hi Chris,

in short - yes. if you remove the SME schedule, then there will still be approx 62GB of data to replicate using snapmirror, if thats the daily rate of change for your log files. If your snapmirrors only update once every 24 hours (daily), then you will have all the data to snapmirror. If say you peformed 2 snapmirrors a day, the amount to transfer would be less.

Guessing this an Exchange volume as you are using SME and that SME is peforming a snapmirror update after a backup operation.

Why not just disable the VSM (volume snapmirror) schedule and just allow SME to control the snapmirror updates..or alternatively disable this in SME and use VSM to control the transfers.

if you use VSM, you can specify a throttle on the transfer and this may alleviate your WAN link issues.

Martin

BOOSTMR2VA
2,283 Views

Thank you for the info, I'm glad to have confirmation on the size of the data.

As far as using SME or VSM, I don't think it matters in this situation.  If I disable one or the other, I'm only saving 2 Gigs.  And even though I do not see a need to have the SME snapshot on the DR system, it's not much extra either way, and may be useful at some point, if the snapshot is a later copy of what the volume snapshot is.

Also, I'm not flooding the WAN link, as the netapps are throttled properly; in fact I have them scripted to change throttles depending on time of day.  At night they run full throttle, during the day they get throttled to half my WAN bandwidth.  I have calculated that with these settings, in 24 hours I can replicate ~200GB across the WAN link.  So I have to make sure my amount of changed data that is replicated is less than or equal to that.  Otherwise the snapmirrors lag behind and start to pile up, causing volumes to become larger and even run out of space.

Public