napp-it zfs volume replication and autosnap

[danishalke]

I have setup replication between two omni napp-it servers (omnios1 and omnios3).
omnios1 ---local server
omnios3 --- Remote server

i want to replicate snapshot of Pool1\zfs1\vol1 to omnios3 and create autosnap every 10 hours and send incremental backup to omnios3.

I tried but autosnap to create snapshot after 15 minutes to test but its not working. it does not create any snapshot and same case on omnios3. it does not pull any snapshot from omnios1.

kindly tell me what are the steps to configure this ?
Thanks

 

[gea]

regarding autosnap:
- have you enabled autoservice (menu jobs-autoservice)?

regarding replication:
if you are using the replication extension, you must:
- create a group (menu extensions - appliance group) on the target server
- create a replication job on the target/backup server

replication jobs and snap jobs are different tasks.
replication creates its own snaps

 

[danishalke]

regarding autosnap:
- have you enabled autoservice (menu jobs-autoservice)?

regarding replication:
if you are using the replication extension, you must:
- create a group (menu extensions - appliance group) on the target server
- create a replication job on the target/backup server

replication jobs and snap jobs are different tasks.
replication creates its own snaps

Thanks ... auto snap is working now... Can you please tell me more about replication steps.
Should i replicate a whole pool or whole ZFS filesystem or just volume. Thats a little confusion for me. ?
for example if want to replicate pool1/zfs1/vol1 from omnios1 to omnios3.

step:
i create a snapshot using auto-snap service on omnios1 and use the following command to send it to omnios3.

zfs send pool1/zfs1/vol1@daily-1393487325_2014.02.27.09.00.03 | ssh omnios3 zfs recv pool1/zfs1/vol1

Then I create o replicate job on omnios3 to pull only incremental data from that snapshot...

is there anything else that i missed ?
Do i have to have new snapshot on omnios1 before replicate to omnios3 each time. ?

Thanks appreciate your help !!!!

 

[gea]

You can replicate a pool recursively or single filesystems/volumes (what I would prefer).
If you want to replicate manually, read Sending and Receiving ZFS Data - Oracle Solaris ZFS Administration Guide

 

[danishalke]

You can replicate a pool recursively or single filesystems/volumes (what I would prefer).
If you want to replicate manually, read Sending and Receiving ZFS Data - Oracle Solaris ZFS Administration Guide

Hi

But on the remote sever.. it is not recommended to create a replicate job reuesively... they recommend to create multiple jobs... one more thing... i am also using RSF-1 High availibility for failover between server1 and server2 and replicate to server3. how can i move snapshot service to server2 in case of failover.

server1 and server2 are in high availibility cluster. pool1 that toggel between these sever in case of one break down..

 

[gea]

Hi

But on the remote sever.. it is not recommended to create a replicate job reuesively... they recommend to create multiple jobs... one more thing... i am also using RSF-1 High availibility for failover between server1 and server2 and replicate to server3. how can i move snapshot service to server2 in case of failover.

server1 and server2 are in high availibility cluster. pool1 that toggel between these sever in case of one break down..

I would also prefer single filesystem transfers.
But I do not really understand your question as the snapshot service is different from replication.

If you ask if you can replicate the same filesystem from zwo different sources at the same time them yes but only if you replicate it to two different target filesystems or if you use a replication script that can use the same snaps and run at different times (or only one is activated at a time). Beside that there is no HA in replications with zfs send.

If you do not use the build-in replication extension in napp-it where you may manually set the same job-id for both jobs so they can use the same snaps for two replications from different source server with same pool or an extra logic that actives/deactivates replication jobs based on any event you must create the needed snaps manually (autosnap service does not really help) or you must script it yourself or modify one of the available free scripts.

Other option is a filebased replication with rsync.