Re: [Networker] Error reported when recovering data?
2011-01-20 23:07:50
Preston de Guise wrote:
Hi George,
On 21/01/2011, at 14:48 , George Sinclair wrote:
Using Linux RH with 7.5.1. I used nwrecover to recover a directory. Everything
appears to have been recovered, but the end of the output reports:
Received 31069 file(s) from NSR server `server'
Recover errors with 1 file(s)
Recover completion time: Fri Jan 21 00:35:48 2011
recover command completed unsuccessfully on Fri Jan 21 00:35:48 2011
Also, it pops up the following error dialog box:
##########################################################################
The recover command completed with an unsuccessful exit code. View the command
output for more information as to why the command failed and possible ways to
correct the problem,
Exit code: 1
###########################################################################
I tried again using saveset recover, and basically the same thing happens.
I did save the output to a file. The output is lengthy, so I sorted it, and I
could see a line near the bottom concerning a file that grew during the save. I
then opened the file in an editor and searched for that line and part way
through, I see this:
./monitor/logs/monitor.log.lck
./monitor/logs/monitor.log
32227:recover: `./monitor/logs/monitor.log' grew by 2585 bytes during save
./monitor/logs/
I also see the following from the savegroup completion notification:
client: /home/log level=full, 3728 MB 00:15:35 32624 files
* <WARNING> : Warning - `/home/log/monitor/logs/monitor.log' size grew
during save
* <WARNING> : Expected 261118187 bytes for
`/home/log/monitor/logs/monitor.log', got 261120772 bytes
1. Can I assume that this is in fact the error that the recover command was
reporting?
Yes, that's the most likely scenario.
The output is so lengthy that checking assiduously line by line would
take a long time, so I just sorted it, and that one popped right out.
There are no references to anything that 'shrunk' or 'changed' size, so
I think the 'grew' line is the only one related to any kind of warning
or error that I can see.
I have seen these messages for sundry files for other backups, and that's OK,
but I didn't think that the recover tool would complain about files that
changed during a save? Is that the normal behavior for both browsable recovers
and saveset recover?
From memory it was introduced somewhere in the 7.x series, thought it may have
been there earlier and I just didn't notice. It should be considered normal
behaviour. Effectively, recover is trying to warn you that some files restored
may not be consistent/correct because they weren't backed up that way.
Yes, that makes sense.
2. How does NW know that a file grew during the save if you're doing a saveset
recover? Where is that information stored? It's not reading the index during a
saveset recover, right?
Presumably it would be part of the saveset stream on the media itself, as I
agree, it wouldn't be consulting the index in this scenario. I would surmise
that something is recorded in the saveset stream to note that the expected file
size was X but it got Y.
Sounds reasonable.
3. I guess NW walks through the directory and takes a look at the file sizes
before the backup actually runs and so it knows that a given file's size has
changed just before it backs it up or maybe while it was backing it up? Not
surprising for log files. I could use logasm, but that only suppresses the
warning, right? I don't really care about the warning. Just wanted to find out
what the error is.
No, NetWorker does not do a pre-backup walk. It walks as it backs up. So the
scenario would be:
walk
encounter file that needs to be backed up
note meta-details about file
backup
compare meta-details about file
warn if changed
continue walking
Ok, got it.
logasm isn't so much about suppressing the warning - it does that as a side-effect, but
it's effectively you telling NetWorker "I don't care if this file changes during
backup".
What does NW do differently when you have that directive in place other
than not issue the warning and maybe not record it in the stream, and/or
index, wherein a recover would not complain like mine did?
Does logasm do more than just that? It doesn't enable the log file to be
more reliably backed up does it?
Thanks!
George
Cheers,
Preston.
--
Preston de Guise
http://nsrd.info/blog NetWorker Blog
http://www.enterprisesystemsbackup.com "Enterprise Systems Backup and
Recovery: A corporate insurance policy"
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the body of the email. Please write to networker-request
AT listserv.temple DOT edu if you have any problems with this list. You can access the
archives at http://listserv.temple.edu/archives/networker.html or
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. -
To sign off this list, send email to listserv AT listserv.temple DOT edu and type
"signoff networker" in the body of the email. Please write to networker-request
AT listserv.temple DOT edu if you have any problems with this list. You can access the
archives at http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER
|
|
|