Networker

Re: [Networker] Problem with groups not starting

2006-08-16 03:27:45
Subject: Re: [Networker] Problem with groups not starting
From: Stuart Whitby <swhitby AT DATAPROTECTORS.CO DOT UK>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Wed, 16 Aug 2006 08:14:28 +0100
Check your scripts that modify this.  The group attribute should be "autostart: 
Enabled" and NOT "autostart: enabled".  I've seen the same with drives.  They 
need to be "enabled: Yes".  Looks like some resource attributes are case 
sensitive and some aren't.
 
Cheers,
 
Stuart.

________________________________

From: Legato NetWorker discussion on behalf of Groth, Jonathan A
Sent: Tue 15-Aug-06 17:35
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] Problem with groups not starting



Denis wrote:

> I've experimented this behaviour (under solaris, too) : no start, no
messages in daemon.log, savegrp.log, savepnpc.log, and of course,
nothing displayed !
>
> Check if there are files under /nsr/tmp, especially if they are named
your-group-name.res.lck.
>
> If so, stop Networker client, delete this files, and restart.
>
> In a normal conditions (i.e no save running, or no other networker
traffic), all files and directories under /nsr/tmp can be deleted (while
networker
> is stopped, of course !)
>
> These lock files (*.lck) are used by networker to verify if there are
groups currently running when attempting to start one.

Denis, thanks for the suggestions.  There are indeed lck files on the
clients, though none with the group names.  I forgot to mention that the
clients are running OpenVMS... I'm pretty much a total newbie when it
comes to VMS, so I don't know if lock files would be created
differently, or if this in any way makes a difference with the groups
not starting.

I will be running a series of tests in the next couple weeks after
deleting the lock files, seeing if I can get some groups to not start,
comparing when lock files are created, etc.


Robert wrote:

> And if you do a command line savegrp  (not autostart) from cron will
> the problemed groups run?  If not, what is the error?   We've
encountered
> an older bug in 7.1.1 and get an error after a command line savegrp of
of the form "RAP root on backupserver does not have configure networker
> priveleges".   This errot is on the std. error output after the
savertp;
> it is not logged in the daemon.log.
>
> It does sound like your autostart update is not working though.

Robert, the problem groups were running fine (for the 6 months since
we've been doing it) before we started using the script... also, the two
groups that didn't run one day, ran fine the next day, with no change or
intervention on our part.

I have not been able to test from cron, tho we did turn RAP logging on
after the problems this month, so maybe we will see something this time!
And I've tested running savegrp from the commmand line without a
problem...

Thanks for reading,

-Jon

-----Original Message-----
From: denis.mail.list AT free DOT fr [mailto:denis.mail.list AT free DOT fr]
Sent: Monday, August 14, 2006 8:02 AM
To: Legato NetWorker discussion; Groth, Jonathan A
Cc: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] Problem with groups not starting

Selon "Groth, Jonathan A" <jonathan.groth AT EDS DOT COM>:

> Hello all,
>
> I'm having an odd problem with groups not starting.  Not only not
> starting, but not producing any messages at all.  I'll put in some
> background info below, but first, let me give the basics:
>
> NetWorker Server:  7.2.1
> OS:  Solaris 8
> Group's Autostart set to Enabled
> Schedule set to proper level (full) on the days in question
>
>
> Background (probably too much):
>
> Customer requires end-of-month backups to start at different times on
> different days. So, a couple ksh scripts were written to set the start

> times in advance, so that people wouldn't have to log in on weekends
> when end-of-month falls on those days.  First script takes input
> (group name, new start time, date you want the change to take place,
> the time the change takes place), then it creates a "script" with all
> the details nsradmin would need to do the deed, and finally it appends

> the input and the location of the nsradmin "script", all on one line,
> to a (what I
> call) control file.  Second script is run from cron * * * * * and
> checks that control file; when the change date and time match, the
> nsradmin "script" on that line is executed and voila! the group's
> start time is changed.  This works perfectly, from what I can see.
But...
>
> ... sometimes (but not most of the time!) the groups that had their
> start time change wouldn't start!  At all!  No messages, nothing in
> the daemon.log, etc.  The times were successfully changed, sure, but
> when the time came, NetWorker just glided past it, doing nothing.  As
> I wrote above, the Groups were enabled just fine and the schedule
level set.
> Also worth noting that the change time was set hours before the backup

> was scheduled to start, so that wasn't a conflict.  Also, this problem

> has never cropped up except in those groups changed by the script. 
> But, only 2 of the 20-30 group changes exhibited this behavior.  The
> rest ran to expectations...
>
> We've talked to EMC and they see no reason why this is happening,
> though they believe its caused somehow by the script(s).  I'm inclined

> to agree, given that this has appeared only when the script changes a
> group, but I would expect it to do the same to all of the groups
> changed, not a small percentage of them.  Its arbitrariness puzzles
me.
>
> Any thoughts would be appreciated.  I can post/email my
> uglier-than-sin scripts, output, whatever, should anyone desire them.
>
> Thanks for reading!
>
> -Jon
>
>

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
wit 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



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
wit 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