cancel
Showing results for 
Search instead for 
Did you mean: 

Catalog replication to DR site and recov ery

liuyang
Level 6
Partner Accredited

Hi, my master and media servers are all NBU 7.0.1 on Windows 2008 R2. We configured that two of the media servers are functioning as deduplication nodes. One will be in HQ site and other will be in DR site. We use SLP to replicate data from HQ site to DR site. We are going to have some DR practice soon. I am wondering how to replicate and recover Catalog data. Currently we only back up catalog data to tape in HQ site. My questions are:

1. Can I replicate Catalog data to dedupe pool in media server in DR site over WAN? Using SLP or Vault? Someone mentioned we can set to back up Catalog data to dual locations. I think this refers to SLP, is it?

2. Assume my HQ site is down, I will build a new master server and recover the Catalog data. According to my friends, it may take days to do that. But one friend mentioned we can just restore Catalog data similiar to normal data. We only need to restore \netbackup\db file to the new master server. This kind of restore is much faster. Is it correct? Does anyone do something similar before?

Thanks in advance.

7 REPLIES 7

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified

Hi,

 

If you want to replicate the catalog you can use a 3rd party hardware replication method like SRDF, or you can use NetBackup Realtime to replicate the catalog (its free). See the attached document.

 

Alternatively, if you're fine with a little longer recovery process you can simply restore your catalog media at the DR site. Install the master server at DR (same name, same netbackup version and patch level), then just run the recovery wizard to recover the catalog.`

 

Its not a good idea to place your catalog on dedupe (and probably not supported)

liuyang
Level 6
Partner Accredited

Hi Riaan, thanks for your replies. I found there is an option to set multiple destination for Catalog. I am thinking to set two destination: one volume pool + one disk storage unit (in DR media server). So Catalog will back up to dual destinations. Is it a possbile solution?

watsons
Level 6

Take a look at this: https://www-secure.symantec.com/connect/blogs/netbackup-71-features-auto-image-replication

This feature uses SLP. Of course it's quite new and I haven't used it, see if you want to test it out and share more :)

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified

Hi,

 

That should work but how big are your links to DR?

liuyang
Level 6
Partner Accredited

Hi Watsons, thanks for the link.

Hi Riaan, my WAN link is 2M leased line. Not sure if it is good enough to replicate Catalog over this link. Is there a way to estimate the size of Catalog data?

AbdulRasheed
Level 6
Employee Accredited Certified

Hi Liuyang,

  Duplicate to remote master (formerly known as A.I.R) is the feature designed for this specific use case.

  Once you implement this feature, you do not need to replicate catalog at all. The meta data for your backups are sent to remote storage server along with backup image so that it can be instantly imported at the remote master server. As your friends had said, preparing a master server at the DR site after an unfortunate disaster does take some time to setup. Or you need to provide a hot standby which is a waste of precious processing resources and energy.

  By using remote to master server feature, your offsite has its own master server which could potentially be backing up clients at offsite. You can replicate from HQ to DR and vice versa at the same time. Right now, A.I.R is supported for Media Server Deduplication Pools which is what you already have. Once you upgrade your environment to 7.1, you have a great DR solution handy.

  See the FAQ on this topic here: http://www.symantec.com/docs/TECH154323

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified

You estimate the size of the catalog using this calculation.

 

150 Bytes * the amount of files backed up * amount of retentions you keep (i.e. if you run a weekly and keep it for 4 weeks, the retention amount is 4, if you run a daily 6 times a week and keep it for 2 weeks, thats another 12 retentions)

 

To work out the daily change you can take the amount of files backed up per day * 150 bytes.