Backup Central

I know I've asked this question before, but it's been a while, so I
can't recall the answer, but just need to clarify.

If you have a save set that fails an incremental then will the next
incremental ignore that one and simply base everything on what's changed
since the previous one (the one before the failed one)?

This is a really big save set, and I'd prefer not to have to run another
full or a numeric unless I have to.

I'm curious if I can rely on the next incremental to really be what it
would have been if, say, the previous failed one had never run.


The save set failed with this message in the savegroup completion e-mail
notification:

--- Unsuccessful Save Sets ---

* client1:/disk1/data 1 retry attempted
* <SEVERE> : remote exec problem for command `save -s server -g
group1 -LL -f - -m server -t 1320909810 -o
RENAMED_DIRECTORIES:index_lookup=on;BACKUPTIME:lookup_range=1317798865:1320909810
-l incr -q -W 78 -N /disk1/data /disk1/data': Connection refused

Cloning is enabled for the pool (pool1), but in the case of the failed
incremental, no clones were made (copies=1).

If it helps, here's a list from mminfo for the save set (/disk1/data),
beginning with the day before the failed incremental:

mminfo -s bootes -q 'savetime>11/10/11,name=/disk1/data,pool=pool1'
-xc'|' -ot -r
'name,savetime,sumsize,totalsize(4),sumflags,ssflags,clflags,copies'
name|date-time|sum-size|total|flags|ssflags|clone-flags|copies
/disk1/data|2011-11-10|72 GB|72 GB|cb|vF||2
/disk1/data|2011-11-11|196 GB|196 GB|ca|vrEiF|E|1
/disk1/data|2011-11-15|561 GB|561 GB|cb|vF||2
/disk1/data|2011-11-16|4 B|4 B|cb|vF||2
/disk1/data|2011-11-17|4 B|4 B|cb|vF||2

I thought when NW runs an incremental for a save set, it looks at the
date/time of the previous incremental as it appears in the media
database, and then any files whose modtime (actually, change of file
status time) is later than that will then be backed up, but if the
previous incremental was aborted then it would instead look at the one
before that, and so on and so forth. And the client file index does not
get read unless a recovery is run, never mind the fact that entries were
undoubtedly made in the client index since 196 GB were backed up before
the save set failed.

I can remove the aborted save set entry (ssid) from the media database
(nsrmm), but is there any harm leaving it in there? I'm not concerned
about the additional space being used in the CFI. That will get cleared
out over time.

Thanks,

George

--
George Sinclair
Voice: (301) 713-3284 x210
- The preceding message is personal and does not reflect any official or
unofficial position of the United States Department of Commerce -
- Any opinions expressed in this message are NOT those of the US Govt. -


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

As NW 'calculates in save sets', it must restart the backup of all unfinished save sets from the very beginning. This happens when you restart a group.

And do not forget: if your backup media would be an AFTD, the aborted save set will immediately be deleted. So there would not be anything left as a reference.

On 2011-11-18 13:41, bingo wrote:
As NW 'calculates in save sets', it must restart the backup of all unfinished save sets from the very beginning. This happens when you restart a group.

Just to be clear, I didn't restart the group using the 'Restart' feature
under nmc (or nsradmin), and within the 'Restart window' value, to force
only the uncompleted save set to rerun. Instead, the group ran again the
next day under it's normal schedule. But is sounds like what you're
saying is that it wouldn't have mattered as far as that save set is
concerned. In either case, NW can't base a backup on an aborted save
set. Instead, it will back up the same files that it would have if the
previous failed save set had never been run, including, of course,
anything else that's been added or changed since then. That right?


And do not forget: if your backup media would be an AFTD, the aborted save set will immediately be deleted. So there would not be anything left as a reference.

That makes sense. But as far as tape is concerned, is there any concern
over leaving the aborted save set entry in the media database? Keeping
it in there allows me to reference it for reporting purposes in the
future. I would never try to recover from it (e.g. saveset recover),
however. I'm not concerned about freeing up index space created by that
save set.

Thanks.

George


+----------------------------------------------------------------------
|This was sent by carsten_reinfeld < at > avus-cr.de via Backup Central.
|Forward SPAM to abuse < at > backupcentral.com.
+----------------------------------------------------------------------


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER



--
George Sinclair
Voice: (301) 713-3284 x210
- The preceding message is personal and does not reflect any official or
unofficial position of the United States Department of Commerce -
- Any opinions expressed in this message are NOT those of the US Govt. -


via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER

...
Instead, it will back up the same files that it would have if the previous failed save set had never been run, including, of course,
anything else that's been added or changed since then. That right?

Correct.

----------------------------

...
That makes sense. But as far as tape is concerned, is there any concern over leaving the aborted save set entry in the media database? ...

No - NW will not touch it by default. If you want to be 100% sure you may of course delete the save set from the media index with "nsrmm -d -S ssid[/cloneid]". This can not touch the backup data but you will loose all references. Only scanner would be able to "re-find" it.