Media DB error

Question

Hi all,

we are trying to recover the index of stor machine from 09/21/2006 by nsrck -L7 -t "09/21/2006" stor but after this we try to browse by NetWorker User for that time and doesn't show any data. We have analyzed the daemon.log and have seen the error

nsrmmdbd: Media DB error: Save set eligibility can not be cleared, no clone instance is within recyclable time period

as you can see in the following fragment of daemon.log:

09/07/07 11:02:27 nsrd: index notice: The file index for client 'stor' will be recovered.
09/07/07 11:02:27 nsrd: server1.w.csn.es:root browsing
09/07/07 11:02:28 nsrd: media waiting event: waiting for sdlt320 tape CSN122 on server1.w.csn.es
09/07/07 11:02:30 nsrmmd #5: Start nsrmmd #5, with PID 22778, at HOST server1.w.csn.es
09/07/07 11:05:04 nsrd: server1.w.csn.es:root browsing
09/07/07 11:05:04 nsrd: server1.w.csn.es:root done browsing
09/07/07 11:05:48 nsrd: index notice: The file index for client 'stor' will be recovered.
09/07/07 11:05:48 nsrd: server1.w.csn.es:root browsing
09/07/07 11:05:48 nsrd: Operation 42 started : Load volume `CSN176'.
.
09/07/07 11:05:49 nsrd: media waiting event: waiting for sdlt320 tape CSN176 on server1.w.csn.es
09/07/07 11:05:51 nsrmmd #6: Start nsrmmd #6, with PID 22816, at HOST server1.w.csn.es
09/07/07 11:06:04 nsrd: /dev/rmt/1cbn Verify label operation in progress
09/07/07 11:06:22 nsrd: /dev/rmt/1cbn is now write protected
09/07/07 11:06:22 nsrd: /dev/rmt/1cbn Mount operation in progress
09/07/07 11:06:22 nsrd: media event cleared: confirmed mount of CSN176 on /dev/rmt/1cbn
09/07/07 11:06:22 nsrd: server1.w.csn.es:index:stor (3/24/06) starting read from CSN176 of 280 MB
09/07/07 11:06:26 nsrd: [Jukebox `STK-L180', operation # 42]. Finished with status: succeeded
09/07/07 11:09:49 nsrd: server1.w.csn.es:index:stor (3/24/06) done reading 280 MB
09/07/07 11:09:49 nsrd: server1.w.csn.es:root done browsing
09/07/07 11:09:49 nsrmmdbd: Media DB error: Save set eligibility can not be cleared, no clone instance is within recyclable time period

09/07/07 11:09:49 nsrmmdbd: Media DB error: Save set eligibility can not be cleared, no clone instance is within recyclable time period

09/07/07 11:09:49 nsrmmdbd: Media DB error: Save set eligibility can not be cleared, no clone instance is within recyclable time period

09/07/07 11:09:50 nsrd: index notice: completed recovery of index for client 'stor'

I forgot to tell you that Backup Server has installed NetWorker 7.3.2.

Thanks and regards.

dpinink_silva · Accepted Answer

Hi Sergio,I saw in the log that you requested index from 09/21/2006, and Nw got from 3/24/06. If index from 09/21/2006 was no longer available, I can imagine that it expired, and so the same happened to the ones before that day. Maybe Nw tries to get the closer to it on the past, maybe 3/24/06 was the closer one before that day, because somehow it got a different retention or was in a tape that in any way was not expired.Do you have any available index for that client before 09/21/2006? If the one from 3/24/2006 was the only one, it confirms what I said. If not, then someone else can help us here

dpinink_silva · Answer

Do you know if the data you want still is in media database? In other words, the savesets still exist in any tape?You could try to run scanner -i on tape and see if that helps.

smora1 · Answer

Thanks Daniel,We found a way to recover what we needed through scanner -i some days ago. As you are saying, the index for requested date no longer is available in tape but my questions are the following:1) Why isn't NetWorker able to detect this? 2) Why does it recover the index of 03/01/07 when there is index closer to the chosen date? I remember in the 7.2.x version this mistake didn't happen.

ble1 · Answer

What about data saveset? There is different procedure for that in 7.3.x and that has been documented now in release notes and has been discussed here too.

dpinink_silva · Answer

Hi Sergio,If in 7.2 it did not happen, I can only believe that this changed in 7.3. EMC may think that the message in the daemon.log is enough to alert you that Nw is restoring the index from another date then the one you requested.Or they may began to act like Microsoft and are trying to create programs that know what is best for you, without asking you

smora1 · Answer

OK but what this error message meansnsrmmdbd: Media DB error: Save set eligibility can not be cleared, no clone instance is within recyclable time periodIt seems that after recovering 03/21/06 index, this index is purged because its recyclable time period has already expired. Could it be?

ble1 · Answer

Actually the first question is if your saveset is expired or not. As for index date, it means that closest index to that date is the one from 24th for that client.

smora1 · Answer

Yes, the index saveset from 09/21/06 is expired.

smora1 · Answer

OK Daniel,I agree with you, the only thing that I don't understand is why NetWorker shows this error message and recover the closer index in the past instead of recover nothing and shows the reason of that.Thank you very much for your collaboration.Obrigado

ble1 · Answer

Yes it could - as I said there is a modified procedure documented and discussed earlier (now you have extra step with nsrmm before restoring index files). Give it a try. As for error message, I have seen it once or twice, but can't remember under what circumstances... could be also index restore. I would advise to use this new method/approach first.

smora1 · Answer

One last question, where is that procedure?

ble1 · Answer

As I said few times, in release notes and KB. Hint: see page 6 and something called 'Changed procedure for recovery when all cloned instances of a save set have expired and no records of the save set remain in the media index'.

rhornsby2 · Answer

I've got exactly the same thing but the procedure didn't work? Am i doing something wrong

Customer tried to run nsrck -L7 on an old date and yes the index for that file has expired.
We identified the index saveset and ran through the routine below - we also did it on the saveset we were trying to recover from.
Once we ran through the routine we run nsrck -L7 -t but again the index recover is only the nearest date to what you requested that is still within its retention period

Should this procedure make the index browsable again or just allow saveset recovery?

Changed procedure for recovery when all cloned instances of a save set have expired and no
records of the save set remain in the media index
With NetWorker release 7.3.2, when all cloned instances of a save set have passed
their retention period and there is no reference to the save set in the media index,
the following procedure must be used to mark a save set as eligible for recovery:
1. Use the nsrmm command with the -e time option to change the retention time
for the clone save set:
nsrmm -e time -S ssid/cloneid
If the cloneid is not identified with the -S option, the following error message
is displayed:
Save set ssid cannot be marked as notrecyclable. Please specify
the ssid/cloneid of the particular clone instance.
2. Use the nsrmm command with the -o notrecyclable option to instruct the
media database that the save set is no longer expired:
nsrmm -o notrecyclable -S ssid/cloneid
If the -o notrecycable option is used with nsrmm prior to changing the
retention time as described in step 1, the following error message is displayed:
nsrmm: Save set ssid:ssid cloneid:cloneid eligibility can not be
cleared, retention time must be adjusted first
After this procedure has been completed, the save set is recoverable.
Note: This procedure only applies if you want to insert expired NetWorker client file index
backups.

ble1 · Answer

That procedure works with data ssid only. If index ssid is expired you will need to scan ssid.

NetWorker

Media DB error

Was this post helpful?