Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations SkipVought on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Re-using tapes. 3

Status
Not open for further replies.

lahrg

MIS
Jun 15, 1999
2
0
0
US
We use Solaris 2.6 and a Sun L280 tape changer. I am having difficulty getting Legato to recycle the tapes at the end of a one week period. I have my browse time and retension set to one week. I have 6 tapes in the changer for backups and I expect it to use the tapes until they are full and then recycle the oldest tape. Am I missing some setup to make this happen? I was under the impression that Legato would handle this automatically. I get about 23 days worth of saves done with the 6 tapes I use.<br>
<br>
Thanks
 
As I understand Networker, the recycle option will only come into effect when *ALL* savesets are recyclable. So if you have more than one backup on a tape taken over several days it will only become recyclable one week after the last saveset backups browse period.
 
If all of the save sets are recyclable and it still does not recycle, in the GUI, click on Media ----&gt; Jukeboxes and pick the library. Then click to view details. At the very bottom of the screen is an option to set Auto Media Management. Set that to &quot;Yes&quot;. That will make Networker automatically recycle. You can also recycle any given tape manually.
 
My understanding was that you could not use auto media management with jukeboxes - only with single tape drives.&nbsp;&nbsp;Whenever I have tried to check that box on our seven tape jukebox properties page, I get an error
 
Nope.... we have a 8 drive tape library and we have the auto media management option turned on.&nbsp;&nbsp;Otherwise, wouldn't have to always check through the pools and manually re-label the tapes that are no longer needed.
 
We have a 10 drive jukebox with 550 slots. If auto media management was off it would become very hectic.<br><br>auto media management looks for new tapes to use when needed, if it cannot find any, it then looks for expired tapes.<br><br>Once it finds what it needs, if auto media management is on, it will relabel the tape, and begin writing to it.<br><br>if auto media management was off, tapes would have to be relabeled manually.<br><br>BW
 
I have encountered this problem as well, we use a 326 slot jukebox, I think what you may have been told, is that you should not use the &quot;AutoClean&quot; feature that resides within Legato, use the autoclean on the hardware, even Legato will suggest that.

CF
 
There are a lot of different threads about this problem and I still haven't got to the bottom of this problem with Networker and I now think its a real bug. We have a Sun E250 running Solaris 8, NW 6.1.1 and an L1000 autochanger. Auto media management is on. We have weekly full backups and a retention and browse period of 4 weeks. Even after 8 weeks tapes do not automagically become recyclable and we have to manually relabel them. They are barcoded. Absolutely *ALL* savesets have passed their browse and retention periods and the tapes stubornly remain full and non-recyclable. This is a very irritating bug. Even Legato have been unable to supply a fix.
 
This is not the norm. The above response are correct.

List the saveset on the volume along with the ssflags

E in the ssflag field represents a saveset that can be recycled.

Have you recently reduced/increased a browse or retention period. Do all clients in a pool have the same bowse/retention period
 
Auto Media Management is only for labeling new media. It does not have to be &quot;on&quot; for tapes to be recycled. You may be having issues with not configuring pools to recycle to other pools. We have issues with forcasting nodes messing up our retention policies because the retension is base on savetime which comes from the clients. Since these node have advanced clocks My tapes don't get recycled. You can build a script base on an mminfo command like

mminfo -r volume -q&quot;volaccess< 7/01/02 ,!volrecycle&quot;

and send that to &quot;nsrmm -y -o Recyclable&quot; in some fashion depending on you os

joe
 
lahrg...
You will have to check the volumes who haven't been recyclable as they should, probobly you will find a client who have savesets located on that volume who is still browseable.

Why are they still browseable?
Well... maybe you have removed a klient from the GUI but it has still some savesets left on tape (mediadatabase).
OR more likely the client hasn't performed the last full-backup correct and is waiting for the next full-backup to bee performed correctly.

I've been working with Legato Networker support for several years and meet many customers who has complained about this, and every time I've checked their media database I've found old clients or failed backups.....

Auto media management....
Has nothing to do with wheter a tape should be recyclable or not, just if it possible for one pool to recycle (lable) tapes from any other pool or not.
 
A tape will only become recyclable if all save sets on the tape are marked recyclable. That's the basic rule.

But you have to remember that you do not have only client save sets but also the index save sets and bootstraps.

They are not visible for recovery - but when it comes to disaster recovery it is an unvaluable point that you have all the index and configuration files on tape.

Do not recycle manually unless you know for sure you do not need any data on that tape. Better think about adding a few tapes more to the library - if this is not a solution there are a few tricks to force writing of index and bootstrap to a certain pool and so restrict it to a few tapes.

Johanes
 
jkalbus...
When you configure a backup-server you shouldn't use a certain pool for index and bootstraps.... you will get almost all bootstraps and index's on one single tape (they are not so large) and what do you think happends if you loose that tape and your backup-sever crashes? You can't recover the mediadatabase. Ofcourse during all great fuckups you will not notice that until it's too late.....

The backup-server should store it's data on several tapes, one simpel solution is to use a short browse and retention time, shorter then all your other servers, and store the data in a &quot;large&quot; pool, then you will never have that problem.



 
Couple of things.

The tapes will not become recyclable until ALL save sets on that tape have expired, and then ONLY after another full cycle has taken place. I would check that there no other backup's with a longer policy have been written to that pool by mistake.

To do this try :

mminfo -q&quot;volume=[volume]&quot; -avr&quot;name,client,savetime(19),ssbrowse(19),ssretent(19),group,pool&quot;

It won't stop it happening but, if your writting to the tape without realising it this may help you spot the saveset that's being written.

Secondly as &quot;soder&quot; just mentioned if you lost your main tape with all bootstrap and index backup's on it your up the creek without a paddle. We have a bootstrap pool, the best thing to do is simply clone that tape. We find that the easiest thing to do, then no worry's about loosing that tape, you loose both someone gets a P45 in their next pay packet.

Hope some of this helps. Regards
KeefB

[lightsaber]
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top