Scanner question

I really love when admins prepare themselves for these cases in advance

The 'scan needed' flag is indicating that you might have save sets on your backup volume which have been created after the last bootstrap. Consequently, these will be missing in the media index after a DR process. The only method to get their info is that you must search the whole volume for save sets which are not yet in the media db.

So, if you run scanner on a volume, NW will do what you have told him to do ... and consequently will scan the whole volume. Depending on your system capabilities and configuration this can take minutes, hours, days or even longer periods. It will obviously be faster if you just run scanner -m instead of scanner -i (file indexes are not really important and may be scanned later, in case you ever need them).

If you could potentially live with the loss of backup data/save sets, you might of course reset the 'scan needed' flag.

Otherwise you will be able to proceed with all other processes like backups, clones, restores etc. 

The best method will be to configure not just one or two but more devices for a pool. This will distribute the workload and speed up processes in general. But do not forget that - even if you optimize NW (for a specific process - the true values will later vary because most likely the environmental situation will not be exactly the same.

Hello bingo.1,

Thanks for getting back to me.

I'm sorry, but how could I have been more prepared in advance for this ? I followed official Networker documentation (NetWorker 19.4 Server Disaster Recovery and Availability Best Practices Guide - Recovering critical NetWorker server databases p.40 and up) and restored the latest bootstrap using nsrdr -N.

I understand why Networker flags "scan needed" and that's ok. There shouldn't be any backups since the last bootstrap, because Networker services were not running on the server. I decided to run a scanner -i -n just to be safe and also because it's the first time I do a DR with Networker and I would like to learn more about it.

I'm not complaining about doing the scan and I certainly don't want to loose any data by just resetting the flag. I'm fine with letting the scanner runs for days; just wondering if I can resume normal operations (mostly backups) while it is scanning. I think it's a fair question to ask since I can't find the answer in the documentation or on the forums.

About running scanner -m instead of scanner -i, again I'm just following the official documentation (here or the NetWorker 19.4 Server Disaster Recovery and Availability Best Practices Guide - Removing the Scan Needed flag from AFTDs and Data Domain Devices p.42 and 43). I understand the difference between the 2 commands, but then again, I'm following official documentation and learning at the same time.

Thank you.

Do not get me wrong - my statement was really a positive feedback. I really appreciate if the admins learn about the product in advance. And as you know there are many important details to discover.

Besides that, understanding the logic behind the official procedures is even better - otherwise one would never detect the true behavior or even problems/errors. The doc writers only document what they will be told to do so. Of course, scanner -i is the 'more complete' solution but the absolute necessary information is just the save set/volume info which is not part of the file index. And do not forget that scanner -i is useless anyway for save sets with the status recoverable or recyclable - scanner -i will read the file index info (if available at all) but then automatically delete it to adjust the save set status to the one listed in the media index.

Thank youbingo.1 for all the information.

Apologies if I jumped. I'm used to people talking to me with sarcasm...

So, just to be clear, following a nsrdr -N, if all my save sets (either tapes or Data Domain) have a Status of Recoverable, Browsable or Recyclable and I know there has been no backups taken following the latest bootstrap ID, then scanning (either scanner -i or scanner -b) is irrelevant. Is that correct ?

We backup daily on DataDomain but monthly on tapes, which are shipped to a storage facility offsite. Following the nsrdr -N, many of my offsite tapes are flagged as "Scan Needed". Not all of them It would be quite costly to recover them from the site and then scan them. I believe it's safe to remove their "Scan Needed" flag. According to the NMC, all save sets on all the tapes are Recoverable. Am I right to remove the flag or should I really scan them ? And why wouldn't Networker flag ALL the tapes as "Scan needed" and only some of them ?

You just must not forget that the bootstrap will only be generated once every day. So suppose that you have a bootstrap backup run at 22:00 and discover a problem at 08:00 the next day which triggers you to run a nsrdr procedure, then the 'new' ones will not be recovered as they never have been made their way into the media index. On the other hand, if you always would run a media db backup (part of the bootstrap) the NW server would just loop around.

The difference should then be covered by scanning the volume, indicated by the 'scan needed' flag. Unfortunately, the is (yet?) to way to trigger the scanner command to just scan save sets with a time stamp later than the last bootstrap

BTW - there are no recyclable save sets on disk volumes as they will be deleted automatically (in hourly? intervals). and yes - there might be recyclable save sets for tapes as for the highest data security, such save sets cannot be deleted separately. Only the whole volume can be relabeled. However, you can do this on a n AFTD/DD (disk) volume.

Yes, if you know that there are no new save sets on tape, you can easily reset their 'scan needed' flag.