Problem with recycable tapes

[LioriK]

Hi,

I'm using Networker 7.4 SP4 and i have configure the pool of the tapes that the retention policy & Browes policy will be one month, in this group i have ~50 clients, and for some reason after a month the tapes dot not recycel automatically.

can you please tell me where is the problem?

Thanks, Lior.

 

[605]

I assume that there are dependencies that you must take under account. Only if ALL backups (the latest) have expired, NW may overwrite the media.

However, NW will set expired save sets to recyclable but not necessarily relabel the media at that point-in-time. As NW does not know scratch pools, NW will only re-use (recycle) media when it is needed (running a new backup).

 

[LioriK]

How should i fix this problem?

Thanks, Lior.

 

[605]

I just made assumptions. As you did not state why this problem occurred, we would digging in mud.

However, if you know the reason, you should already be able to fix it.

 

[LioriK]

ok, thanks.

 

[johnmoore6]

Hi All,
May I jump into this thread. I am having similar problems with my VTL. When I look at the tape, you are right there are savesets that still have not expired. But then I have tapes that are two, three months old and they still have not expired and like Loirik I have set retention and browse to one month. Is there any way to look at the tape to see what is preventing NW from flaging the tape as recyclable. Thanks.

 

[605]

Use
mminfo -q "volume=name" -r "ssid,sumflags,name"
to have a look at the save set statuses. This should tell you more.

 

[johnmoore6]

Hi 605,
Thanks for the reply, and thanks for the suggestion. Might I add that it also might be useful to add ssretent and ssbrowse to the command, and I put client to help decipher what was what. With this I think that I have found a client or two that still might have their retention time set to more than a month. Thanks, I will go through a few more tapes and let you know.

 

[605]

Browse and retention times are fine but do not give you more information at this point in time. "sumflags" already provides the info about the save set status which is sufficient. I just wanted to check whether there are save sets stored on this media with the key letters "b" (browsable) or "r" recoverable and not "E" recyclable.

Yes, in fact the client is a good idea.

 

[johnmoore6]

I understand the that sumflags will provide that information, but if you have a saveset flagged with "b" or "r" and some that are flagged with "E" then it is logical that the tape is not going to expire until the "b" and "r" savesets have reached their time. But as this has become a persistent problem I wanted to know which client was causing the problem and which value was the one that was set to high. "b" and "r" will not give that you that view. But that is small part of the problem. I have a ssid which is over 800 gb, so the saveset is spread over 7 volumes. The sumflag has values of "hb", "mb", and "tb", yet the browse and retention value shows that the saveset has already expired yet the tapes have not been flagged as recyclable. Any thoughts as to where I might look or am I overlooking something. Thanks in advance and I hope that this msg was readable.

 

[605]

Of course, if you know the client, you will most likely open his resource and look at the browse/retention policies.

Do you want to tell us that the retention date of a save set has passed but that it still browsable? - this looks weird and should not happen. If so, i assume that the index manager process (nsrim) did not run (correctly).

 

[johnmoore6]

Thanks 605, I know after I posted the message that it probably was confusing to read. Thanks for breaking it down. So to answer your question, yes the saveset is still browsable yet it has passed it's 30 day browse and retent policy. You mentioned nsrim potentially not running correctly. What can I take a look at to investigate if that is the case? TIA

 

[605]

First of all, make sure that there are no dependencies. In other words:
- If you have later backups (i.e. incrementals or differentials) which still depend on the (full?) save set you are monitoring, it is obvious that this one backup can not be deleted although it has already expired.
If this is the case, the problem is configuration and/or wrong expectations. Earlier backups can only expire if a new backup cycle has started.
- However, in case of no dependencies, i am afraid that there is another problem.
In this case, try to run nsrim from the command line with more verbosity and see what the outcome is. Worst case, you need to contact support.

 

[LioriK]

Hi,

Can you please what is the correct process (nsrim) to run to fix the problem with index?

Thanks, Lior.

 

[605]

Well, as you can see prom the manpages, there are not so many options. nsrim by itself should already do the job.

 

[johnmoore6]

Thanks for the reply. In fact I ran nsrim -X because I had a couple of crashes of the server. We will see if helped. Thanks.