[Networker] nsrindexd -a ADD

networker server Solaris 6
networker version 6.1.2
Client experiencing problems:
        NT
        W2K
        Solaris 6 and 8


Problem: During backups (Server initiated, Client initiated, Oracle RMAN,
Oracle EBU)
the nsrindexd process hangs on an ADD and a READ.  This does not happen
all the time, but it does seem to happen more on Oracle RMAN and Oracle
EBU.  It also seems to happen more when the backup load on the server is
large,
however I have seen it when only one backup streem is coming into the
server.
We are aware of our lack of power when it comes to our networker servers,
and
it is true, at least with us, that Legato is a CPU hog (0% idol time when
reading
tape labels and positioning).

Backup Server Configuration
*       Sun Ultra Enterprise 3500, Solaris 2.6 with kernel patch 105181-32
(June 2002, newest)
*       2 x 336 MHz  RISC processors, 1.2 Gig RAM
*       Legato server 6.1.2
*       Server parallelism = 4, target sessions = 1 for each DLT7 tape drive
*       1 ATL tape library with 4 DLT7000 Quantum tape drives
*       3 differential SCSI connections between the server and ATL tape
library
*       Gigabit ethernet card



*       Sun Ultra Enterprise 5000, Solaris 2.6 with kernel patch 105181-32
(June 2002, newest)
*       4 x 168 MHz  RISC processors, 1.2 Gig RAM
*       Legato server 6.1.2
*       Server parallelism = 7-8, target sessions = 1 for each DLT7 tape
drive
*       2 ATL tape library with 8 DLT7000 Quantum tape drives
*       3 differential SCSI connections between the server and each ATL tape
library
*       Gigabit ethernet card

The hangs happen randomly, no specific time, no specific
server (the best kind of problem).  The backup really does not fail it just
sits in an
executing status.  I have even seen it happen where the networker server
hangs on
its backup of itself on the index ADD.  My co-worker and I have worked on
this
problem for 3 months, we are both relatively new to legato but have about 12
years
of Unix administration between us.  We have exhausted almost every scenario
that
we can think of and Legato is at a loss as well.  Any info would be great,
or even
if one of you have experienced the same problems, I would like to hear from
you.

Thanks
Nick




--
Note: To sign off this list, send a "signoff" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=