Indexes not being saved in Legato database

Oh, I just read this again - so data disks are returning nothing as well... what happens if you try to run save in debug mode from the client? Or for the start you can ran ordinary save... something like:
save -s solegato2.ctstateu.edu -bVista -lfull -v /SYSO/u07

Do you get same issue for both csuvistadev2.opvt.ctstateu.edu and libctsvr.libraryofconnecticut.org (index and data)? If yes (and even if no) I would first jump to 7.1.4 or 7.2.1 build 314.

Probably not related, but.. - update to last patch level which for 7.1.x is 7.1.4. Second, is it possible that you have no index save for this client?

I can see this is again solegato2 server which already had some index mess in previous topic so... What happens when you execute nsrls libctsvr (should be the same as nsrls libctsvr.libraryofconnecticut.org)? Can you check when was the last time this index was saved via mminfo?

hcrvelin,

Actually, this is a different Networker server, the previous one is solegato1 and this is solegato2. Here's the output for nsrls which show nothing in the database:

solegato2:/root> nsrls libctsvr.libraryofconnecticut.org

/nsr/index/libctsvr.libraryofconnecticut.org: 0 records requiring 0 KB
/nsr/index/libctsvr.libraryofconnecticut.org is currently 100% utilized

mminfo shows nothing was ever saved.

solegato2:/root> mminfo -c libctsvr.libraryofconnecticut.org
mminfo: no matches found for the query

Vic

OK, go to the client (libctsvr.libraryofconnecticut.org) and from there do:
savefs -s solegato2 -pv

Do you see expected mounting points to be shown? Compare this with /etc/fstab. If yes, then pick one and run save:
save -s solegato2 -v bUnix -lfull /

If that fails try modified one:
save -s solegato2 -c libctsvr.libraryofconnecticut.org -v bUnix -lfull /

And if that fails too then debug mode:
save -s solegato2 -D9 -c libctsvr.libraryofconnecticut.org -v bUnix -lfull / > output.lst 2>&1

Before all of that make sure that name resolution is fine just for the case we are missing obvious. Also check if nsr* is running as root.

And of course, update to 7.1.4 at least!

hcrvelin,

Looks like you found the problem with savefs. What it means I don't know. Again, any ideas?

[root@libctsvr root]# savefs -s solegato2.ctstateu.edu -pv
savefs: path /var/spool by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Searching for NetWorker bin 'pathownerignore' file.
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /var by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /usr/local by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /usr by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /tmp by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /home by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path /boot by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
savefs: path / by default belongs to client solegato2.ctstateu.edu and NOT client solegato2.ctstateu.edu!
savefs: Default client index for scheduled save will be that of solegato2.ctstateu.edu.
type: NSR client description;
pools supported: Yes;
migration supported: Yes;
browse time supported: Yes;
multiple balanced streams supported: Yes;
remote user: root;
groups: root, bin, daemon, sys, adm, disk, wheel;
arch: i686;
client OS type: Linux;
CPU type: i686;
CPUs: 2;
kernel arch: i686;
machine type: desktop;
MB used: 4678;
NetWorker version: 7.1.Build.230;
OS: Linux 2.4.21-37.ELsmp;
version: 7.1.Build.230;
save set: path=/, level=incr, time="Wed Feb 8 22:19:14 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/boot, level=incr, time="Wed Feb 8 22:19:11 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/home, level=incr, time="Wed Feb 8 22:19:13 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/tmp, level=full, diskno=0, max_sessions=8, stype=save,\
path=/usr, level=incr, time="Wed Feb 8 22:20:15 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/usr/local, level=incr, time="Wed Feb 8 22:19:12 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/var, level=incr, time="Wed Feb 8 22:20:37 GMT-0500 2006", diskno=0, max_sessions=8, stype=save,\
path=/var/spool, level=incr, time="Wed Feb 8 22:24:40 GMT-0500 2006", diskno=0, max_sessions=8, stype=save ;
parallelism: 8
[root@libctsvr root]#

That explains why we see solegato2 being mentioned when running savefs on the libctsvr. Check /etc/hosts and DNS for response on that:
So, for each (and from each) do:
nslookup hostname
nslookup hostname.domain
They should return same IP.
nslookup IP
This should return FQDN (hostname.domain)

Make sure you have (if you have of course) same state in /etc/hosts. You should test this for both hosts and from both hosts (total of 12 queries and 2 hosts table checks).

One more thing. I just determined that the indexes for libctsvr are going into the indexes database for solegato2. In daemon.log there are entries for the disks being backed up but the client is listed as solegato2. So, somehow libctsvr is being confused with solegato2.

Vic

Hi Mark,

You are right. I did a typo ignoring input of echo command... so what I was trying to say was:
"echo print | nsradmin -p 390113 -i - -s libctsvr"

hcrvelin,

That nsradmin command pointed to the problem. From solegato2 it came back with a line:

name: localhost.localdomain;

Other clients came back with its FQDN. Looked at /etc/hosts on libctsvr and noticed that the entry for 127.0.0.1 looked wrong. It was:

127.0.0.1 localhost.localdomain libctsvr.libraryofconnecticut.org localhost libctsvr

Changed it to:
127.0.0.1 libctsvr.libraryofconnecticut.org libctsvr.libct.org libctsvr localhost.localdomain localhost

Restarted the client and tried nsradmin again. It now comes back with the FQDN. Also, the savefs command doesn't generate those errors anymore. Tried the test backup using your save command and then checked the index database for libctsvr. IT'S THERE!!! So, problem is solved (and it was a good learning experience).

Points awarded. I'd buy you a beer but you live too far away.

Thanks for all your help.

Vic