I have a Networker 8.1.1 server installed on my site with 50 to 60 clients. All backups are going to 1TB SAN partition and after staging moving on physical tapes. I need to resolve following issues; Experts plz respond
Dear experts i would have more questions once i will be succeeded to resolved the aforementioned 3 problems. Your kind advises are awaited.
Here are my suggestions:
1. First you have to check if the cloning is being done by the Staging configuration or by the cloning pools. Keep the disk backup pools and the tape pools separate. for example in you case you using the same pool for backup and as target in your staging configuration. So, analyze where exactly the cloning is configured.
2. Have the target sessions on the device configuration high enough so that it can cater all the sessions that the clone job is sending to the device. If you have a single pool for the daily pool try to restrict just one tape drive for that pool so that all the clone sessions are sent to that tape drive alone.
3. The index backups - index:client_name are required only in case of a disaster recovery. The index files are stored on the server which are used for the browsing the current backups. The option of "No index Save" will skip the index file backup to the backup device only. The index files on your backup server will continue to be on it.
1. I assume that for your Staging resources you have not changed the default setting for the "Destination pool" which is "Default Clone". Please correct that.
2. This assumes that you never have more data than one tape can store. Otherwise, you do not have a choice.
With respect to configuration - i would not recommend to increase the target sessions as this could spread the data too wide which will result in a recover process that takes much longer than expected.
My personal favorite: Script it and stage save set by save set. Then - as the last steps of the script - set the tape to 'full' and eject it.
3. Index backups are NOT necessary for recoveries at all, not even for a DR. You still have save set recoveries in this case.
Index backups are nice for granular (point-in-time) restores.
As file index data is added with each new backup, it makes sense to save the new instance right away. This is what NW does at the end of each automatic (savegroup) backup. "No index save" will skip this process. This solely makes sense, if you run index backups in a different way, (i.e. "savegrp -O"). However you need to be aware that. ..
- you do not have an immediate backup of your CFI database
- you must do this for all clients which are member of that group. "savegrp -l full -O -G groupname" would be the best command to run.
On the last point, i believe the for doing bare metal restore for Windows you would need the index on the backup server else the backup wont show up in the list of backups to be selected in the restore wizard. Is this wrong ?
because I did no expect any technical necessity, I tested the behavior using the latest NW version 22.214.171.124.
And you are absolutely right - I can confirm that behavior. Thanks for letting us know.
Fortunately, the latest backups usually should not have passed their browse date.
I can only imagine that the DR procedure is written in a way that certain files need to be restored first.
Thank you for your informative stuff and apologies from my side for late feed back, actually i was stuck in other tasks.
Finally i am successful with segregation of media pools and clone copy. I have set the threshold of 2% for backed up data. My all backup jobs run at mid night and transfered to physical tapes till morning, then i have configure clone which runs at 10 PM daily and completes approx 600 GB data in about 90 minutes. The next step with cloning is to restrict PTL & NW to touch new cartridge daily for clone job because my clone data is about 600 GB and cartridge remains append able. I have contact emc chat support and they recommend to switch cartridge mode to read-only so that NW never touches that again and always pick a new media for clone daily. I need to automate this process and need help from you guys in this regard.
Just scheduled a job later which will retrieve the tapes written today and set them to full or read-only.
The core lines should look like (Windows)...
mminfo -q "pool=clone_pool,savetime>=today" -r "volume" > tapes.txt
for /F %%I in (tapes.txt) do (nsrmm -y -o full %%I)
Sure we could do that ... although I personally think it is now your turn to verify the suggestions.
It is not too difficult and could be done in minutes. And you have to do it anyway.
Do you have a certain issue where you still need support?