Add volume to VFILER DR

[tomekp]

Hello,

I would like to add new volume do VFILER DR

I know the proceedure, I have done it before (with vfiler dr resync on DR-site to "suck volume into" stopped vfiler).

But here is the trick. The destination vfiler doesn't have two volumes. Someone created volume on source, destination, added volume to vfiler on source...

and created snapmirror session instead of "sucking volume into DR vfiler".

How can i fix that? if I create new one with all the steps and proceed with vfiler dr resync.. what will happen with those two?

• Will NetApp use existing snapmirror relationship (altering the schedule for every 3 mins),
• will it fail because the snapmirror relationship already exist,
• or will it break it and do the baseline copy ?

My first post from this account, greetings everyone 🙂

cheers,

Tomek

[JGPSHNTAP]

First, what version of 7-mode are you using

See if this link helps you out.

https://library.netapp.com/ecmdocs/ECMP1155585/html/GUID-18F85463-10F4-49EA-984C-FCADF4060217.html

[tomekp]

Hi JGPSHNTAP,

thanks for quick answer. But let me do something a little impolite and answer your question with a question

> First, what version of 7-mode are you using

Why does it matter? Will Data ONTAP act differently with vfiler dr resync while using ontap 8.0, 8.1 or 8.2.3  ?

If so, to have a full picture it would be better to know what to expect in each version 🙂

thanks for the link, unfortunatelly it does not provide an answer to my question

There is one sentence

If a volume with the same name exists, the volume is automatically added and initialized for SnapMirror transfers from the disaster recovery vFiler unit. Any existing data on the newly added volume is lost. 

But i'm pretty sure it does not apply to volumes which are already snapmirror destination volumes.

When I thought about it a little, I suspect that vfiler dr resync command will ignore those already existing volumes, due to the fact that snapmirror relationship already exist.

But i would love some confirmation.

[tomekp · Accepted solution]

Since I couldn't find any answer I decided to simulate it 🙂

netapp80 - primary filer;                      NetApp Release 8.0RC3X8 7-Mode

netapp823 - secondary "DR" filer        NetApp Release 8.2.3 7-Mode

netapp80> vfiler status -r

vf1                              running
   ipspace: default-ipspace
   IP address: 192.168.21.85 [e0d]
   Path: /vol/vf1_root [/etc]
   Path: /vol/vf1_vol1
   Path: /vol/vf1_vol2
   Path: /vol/vf1_vol3
   Path: /vol/vf1_vol4
   UUID: f39426fa-9cbf-11e4-80a1-000c29943a5e

netapp823> vfiler status -r
vf1                              stopped, DR backup
   ipspace: default-ipspace
   IP address: 192.168.21.85 [unconfigured]
   Path: /vol/vf1_root [/etc]
   Path: /vol/vf1_vol1
   UUID: 2944999b-839d-4e77-aca6-5d0c2e449866

Additionally to simulate the situation i have created manuall snapmirror relationships for vf1_vol2 and vf1_vol3

netapp823> snapmirror status
Snapmirror is on.
Source                Destination           State          Lag        Status
netapp80:vf1_root     netapp823:vf1_root    Snapmirrored   00:01:59   Idle
netapp80:vf1_vol1     netapp823:vf1_vol1    Snapmirrored   00:02:01   Idle
netapp80:vf1_vol2     netapp823:vf1_vol2    Snapmirrored   00:01:21   Idle
netapp80:vf1_vol3     netapp823:vf1_vol3    Snapmirrored   00:01:14   Idle

Please note, only vf1_root and vf1_vol1 is "sucked into" vf1 (stopped, DR backup)

Now, I have create vf1_vol4 on both sides and do vfiler add vf1 /vol/vf1_vol4 on source

let the test begin! from destination netapp823

netapp823> vfiler dr resync vf1@netapp80
netapp80's Administrative login: root
netapp80's Administrative password:
Volume vf1_vol2 (which is a storage path of remote vfiler vf1)
is already a SnapMirror destination for a source.
Couldn't add /vol/vf1_vol2 to local copy of vfiler vf1@netapp80
Volume vf1_vol3 (which is a storage path of remote vfiler vf1)
is already a SnapMirror destination for a source.
Couldn't add /vol/vf1_vol3 to local copy of vfiler vf1@netapp80
Wed Sep 30 20:43:04 GMT [netapp823:vFiler.storageUnit.off:warning]: vFiler vf1: storage unit /vol/vf1_vol4 now offline.
Volume 'vf1_vol4' is now restricted.
netapp823> Wed Sep 30 20:43:14 GMT [netapp823:vFiler.storageUnit.On:notice]: vFiler vf1: storage unit /vol/vf1_vol4 now online.

netapp823> vfiler status -r
vfiler0                          running
   ipspace: default-ipspace
   IP address: 192.168.21.52 [e0b]
   IP address: 192.168.21.53 [e0c]
   IP address: 192.168.21.54 [e0d]
   IP address: 192.168.21.51 [e0a]
   Path: / [/etc]
   UUID: 00000000-0000-0000-0000-000000000000

vf1                              stopped, DR backup
   ipspace: default-ipspace
   IP address: 192.168.21.85 [unconfigured]
   Path: /vol/vf1_root [/etc]
   Path: /vol/vf1_vol1
   Path: /vol/vf1_vol4
   UUID: 2944999b-839d-4e77-aca6-5d0c2e449866
netapp823> snapmirror status
Snapmirror is on.
Source                Destination           State          Lag        Status
netapp80:vf1_root     netapp823:vf1_root    Snapmirrored   00:01:26   Idle
netapp80:vf1_vol1     netapp823:vf1_vol1    Snapmirrored   00:01:25   Idle
netapp80:vf1_vol2     netapp823:vf1_vol2    Snapmirrored   00:03:48   Idle
netapp80:vf1_vol3     netapp823:vf1_vol3    Snapmirrored   00:03:41   Idle
netapp80:vf1_vol4     netapp823:vf1_vol4    Snapmirrored   00:00:25   Idle

🙂

[scottgelb]

I had this in my training labs (from the Insight conferences) from 2009-2012 when we did the training.. see attached lab05 page 7 for vfiler dr tasks including adding a volume or ip..

[JGPSHNTAP]

Sorry, I didn't respond to you..

now I can comment on your code level since you posted it.  Why on earth are you running 8.0 RC3 X code.    Time to upgrade my friend...

Don't forget to review your snapmirror.conf file after you run resync.

Also, you can tear down and rebuild

[scottgelb]

Good point..one of my pet peeves... save snapmirror.conf before resync... it will update the schedule to every 3 minutes automatically and need to set back after.  I had an RFE to change that but never happened.

[tomekp]

thanks guys,

really helpful!

JGPSHNTAP please note that was just simulate environment with two simulator, i happen to had 8.0 and 8.2.3 on my drive, and I decided to check if they can 'talk to each other for vfiler dr' 🙂

[tomekp]

Since I have my test environment up and ready I decided to do another test

question: what will happen if I break the snapmirror relationship for volume that is not added to vfiler DR. Will vfiler dr resync add this volume to DR vfiler? Will it "resync" the existing snapmirror relationship?

will it erase all data on destination ?

actions:

1. Create a file on source (netapp80) volume

netapp80> wrfile /vol/vf1_vol2/test.txt
this is test
read: error reading standard input: Interrupted system call
netapp80>

2. Update snapmirror and read the file from destination

netapp823> rdfile /vol/vf1_vol2/test.txt
/vol/vf1_vol2/test.txt: No such file or directory
netapp823> snapmirror update vf1_vol2
Transfer started.
Monitor progress with 'snapmirror status' or the snapmirror log.

[ after a minute]

netapp823> rdfile /vol/vf1_vol2/test.txt
this is test

3. break the snapmirror relationship from destination

netapp823> snapmirror break vf1_vol2
Wed Sep 30 21:16:13 GMT [netapp823:wafl.scan.64bit.space.done:notice]: The 64-bit space qualifying scanner has completed running on Volume vf1_vol2.
snapmirror break: Destination vf1_vol2 is now writable.
Volume size is being retained for potential snapmirror resync.  If you would like to grow the volume and do not expect to resync, set vol option fs_size_fixed to off.

4. perform vfiler dr resync from destination and check what will happen 🙂

netapp823> vfiler status -r vf1
vf1                              stopped, DR backup
   ipspace: default-ipspace
   IP address: 192.168.21.85 [unconfigured]
   Path: /vol/vf1_root [/etc]
   Path: /vol/vf1_vol1
   Path: /vol/vf1_vol4
   UUID: 2944999b-839d-4e77-aca6-5d0c2e449866
netapp823> vfiler dr resync vf1@netapp80
netapp80's Administrative login: root
netapp80's Administrative password:
Volume vf1_vol3 (which is a storage path of remote vfiler vf1)
is already a SnapMirror destination for a source.
Couldn't add /vol/vf1_vol3 to local copy of vfiler vf1@netapp80
Wed Sep 30 21:17:22 GMT [netapp823:vFiler.storageUnit.off:warning]: vFiler vf1: storage unit /vol/vf1_vol2 now offline.
Volume 'vf1_vol2' is now restricted.
netapp823> Wed Sep 30 21:17:35 GMT [netapp823:vFiler.storageUnit.On:notice]: vFiler vf1: storage unit /vol/vf1_vol2 now online.
netapp823>
netapp823> vfiler status -r vf1
vf1                              stopped, DR backup
   ipspace: default-ipspace
   IP address: 192.168.21.85 [unconfigured]
   Path: /vol/vf1_root [/etc]
   Path: /vol/vf1_vol1
   Path: /vol/vf1_vol4
   Path: /vol/vf1_vol2
   UUID: 2944999b-839d-4e77-aca6-5d0c2e449866

and now the most important thing. Did I loose my data from secondary filer and got base-line copy?

didn't catch that, since it was only one file ;/ maybe anyone know answer to this question?

[scottgelb]

resync will pick up from a snaprestore where it left off so not a full mirror. However, (see prior lab attachment) if setting up a NEW vfiler dr create for volumes that are mirrored, by default vfiler dr will create new mirrors...not ideal especially with large volumes. But there is a "-u" option that will tell it to use existing mirrors.  But resync always does a snapmirror resync... and used to be our workaround to not remirror prior to the -u option in 7.3.3

[tomekp]

I did another test and actually results are quite different

I have put some data on the remaining vol that was still not added to "vfiler dr"

snapmirror update took about 4 minutes, then I went with

1. snapmirror break {on destination}

2. vfiler dr resync vf1@netapp80   {on destination}

and.. unfortunatelly the baseline copy is transfering, even though nothing had changed on source

So my conclusion is:

If someone made a mistake and create manual snapmirror relationship between filers, instead of vfiler dr resync and let netapp to create snapmirror relationship you can fix it with

snapmirror break; vfiler dr resync but that will cost you full baseline copy for that volume

[scottgelb]

For existing volumes it will resync.  If the mirror was not part of the vfiler dr, then this makes sense... A workaround for this would be to vfiler dr activate the target...then add the volume to the target vfiler.  Then dr resync and it should resync instead of baseline since already part of the vfiler target.  This was the workaround we used prior to the -u option... to create the target vfiler first with the volumes...then resync after with the volumes existing already...

[JGPSHNTAP]

Also, if there is major confusion you can always delete the dr vfiler and recreate it with -u at the end to ensure it doesn't snapmirror the data and just rebuilds the dr vfiler.