Where do you set your retention?

In networker administrator console, in client properties part.

I assume you see this for RMAN savesets?  Perhaps you wish to double check it to confirm that as in the past I have seen when backup of nwora.res is done at the end some different properties are applied.  Further, if your volumes are grouped into pools based on retention (eg. this pool has tapes with data which should all have same retention) you could also push retention on pool which from my past experience (7.4.x) works just fine.  Additionally, you may push retention within RMAN script, but that should not be necessary.

Yes rman savesets. Pools are grouped according to retention I will try pushing it to 1 months, thanks. And I also checked rman script, there is nothing about retention there. Is it possible to update savesets with this new config in order to check it worked ?

Once you set retention on pool it will be picked up by all new savesets onward.  Existing one you will have to manually correct with nsrmm I believe.

Hey,

Forcing retention via pool onfiguration seems not to work. I have noticed an error in the deamon log. Maybe it is reletad with my issue, what do you think?

" 07/12/10 14:39:47 nsrmmdbd: Save set ssid:2621112061 cloneid:1278934789 retention time update is before browse time, adjusting browse time "

How are judging that the retention is not working?  Are the dates not what you expect from:

mminfo -q ssid=[ssid] -r ssid,name,sumflags,savetime,ssbrowse,clretent,ssretent,copies,volume,volretent

If you can provide more detail on what retention policies you have set where and what results you are seeing from mminfo, we should be able to sort this out.

-Bobby

For example one of my rman client that is not obeying retention. It needs to be recyled every 3 months. Only one client is writing to that pool. As you can see its retention is invalid.

I think without really looking into configuration and whole workflow you have you won't get far.  Perhaps you wish to open this with support or simply upload tarball with your configuration here.

Hi Strurm,

I do not wish to sound as vendor support which for any issue will suggest upgrade, but 7.3.x code didn't stay in nice memeory of mine.  It is also unsupported nowadays,

While I run different OS platform, for what is worth, I can say that with 7.4.5.x (which I run) I do not see any issues (at the moment upgrading test environment to 7.5.2.x while rest of environment will be done after summer vacations).  I use 1 single retention (and browse) for ALL clients.  This applies to long-term backups too, but their retention is altered by pool setting.  In my case it works, in your case it sounds as bug which might be linked to NW version.

I'm not sure how much of an option is to upgrade for you, but it would be strongly recommended. Actually, those long retentions you mentioned in first post do sound as something already seens with 7.3.2 and most probably later 7.3.x code too.

For example, I have 2 groups. One is file level and other is rman backup. "Konfyon_app" and "konfyon_db" are their names. These 2 groups are at "konfyon" backup pool with index storage. All clients in these groups have 3 months retention and browse times. All clients are scheduled to be backuped up daily (6 days incremental,sunday full). I have also forced 3 months of retention from pool configuration. This problem is not only for this pool. I have noticed that some other pools are also behaving like this. I am using HP EML245 library with lto3 tape drives on Windows 2003 sp2 server. Networker version is 7.3.3. As you can see in the pictures retention times are set to ../../46 .

Thanks...

Hey,

Thanks for the answer. Yeah no longer supported and we have also ordered a new backup system. But it will take about 2 months to be fully functional I was looking for a temporary solution so that I wont be able to monitor pools every hour.

I'm not sure if you have tried this before, but perhaps specifying NSR_SAVESET_BROWSE and NSR_SAVESET_RETENTION in rman script could do the trick (even I would hold my breath on this one if this is server side issue). 

There is also a way to switch off NW policies with NSR_RETENTION_DISABLED and to rely on Oracle policy, but I never used that before so I can't give much comments on that one.