Back to basics first. Check that the browse policy isn't set by default to
something like "2 minutes" and that the group doesn't have "no index save"
checked.
Thankfully, your server's Solaris. Get hold of a copy of the appropriate lsof
for your OS version and install it. Run a backup of this client from a
savegroup. From the server, run a "ps -ef | grep ADD" to check the PID of the
nsrindexd -ADD process for this client. Now run "lsof -p <PID>" to identify
whether this is saving to a particular file. Otherwise, get in touch with
Support (or maybe just a shout to the list) to get a hold of dbgcommand for
Solaris and run 'dbgcommand -p <nsrindexd's PID> "Debug=9"' (the main
nsrindexd, not any of its child processes), run your group, then run
'dbgcommand -p <nsrindexd's PID> "Debug=0"'. Check the daemon.log afterwards
to see if the main nsrindexd is complaining about anything. Alternatively (or
as well), run the savegrp for this client from the command line using -vvv to
get more output, and then try again using D3 and D9 on the group if nothing
else has helped as yet.
Main thing that comes to mind is that I'd be surprised if this works if it's
not set up to work that way. I don't think this will be firewall issues or
similar preventing the client from talking to nsrindexd on the server, or save
just wouldn't understand and would commit some form of binary suicide (hacks
triangular hole out of its memory and waits for OS to decapitate it). Main
thing to do is to figure out what's configured to save this way.
HTH,
Stuart.
________________________________
From: Legato NetWorker discussion on behalf of Stan Horwitz
Sent: Tue 25-Jul-06 02:50
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Subject: Re: [Networker] client is not getting index entries written...
On Jul 24, 2006, at 7:29 PM, Brian O'Neill wrote:
> Added info:
>
> If I run a manual "save" from the client, and I specify the
> client's name with the "-c" option, the data winds up in the index
> (and it seems to even back-populate the other filesystems as well,
> if that's even possible). However, the default save command run
> from savegrp, which doesn't use the "-c", results in no index data
> being stored.
>
> Why isn't "save" getting the proper name? "hostname" returns the
> short name of the host. I've tried adding the client with both
> short and long name, and it doesn't seem to matter.
I have seen that kind of behavior happen with multi-homed clients and
also with a DNS error. Check both.
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
|