Xerrors on HP-UX 10.20

[Francois Bariselle_3]

Hi all,

I have a K580 HP-UX 10.20 standalone server
with 4 Gb of memory.
Many clients connect on this server with Reflexion X (XDMCP direct connection) to run hp-term and launch rpc programs.

I can see troubles with client pid such as :

- The hp-term take about 30 sec to edit a file with dtpad. (Only for the first time after his login)

- Many programs pids are same for two users! and cause program crash.

- I look in /tmp, /var/syslog and no errors!

- I can see a log file in /var/vue/Xerrors that it contain somes errors since 16 Apr 2004.

1) This is a corrupts users?
2) This is a mismatch with the version of Reflexion X and HP-UX ?
3) If no, any idea?

You can consult the attachment file "Xerrors"

Thanks in advance,

Francois Bariselle, ing
SDMI Technologies

[Mel Burslan]

how many processors does this poos K580 have I am wondering ? Since you said "many users connecting via Reflexion X" I am wondering if the horsepower of the CPU is not enough.

especially when you say "takes too long to launch the dt pad for the first time, I am more inclined to think that the initial setup of the session is taking too long but once it is established and application is in memory, user is fine, is making me want to say, you are having CPU insufficiency issues.

Just my thoughts.

[Francois Bariselle_3]

More information,

Only 8 users use Reflection X to connect on the server.

you can consult the attachment for look the CPUs load.

Perhaps the sar command cand help you ? :

# sar -v 5 5

HP-UX quebec B.10.20 U 9000/800 05/05/04

11:31:40 text-sz ov proc-sz ov inod-sz ov file-sz ov
11:31:45 N/A N/A 333/1620 0 852/1988 0 2439/3219 0
11:31:50 N/A N/A 333/1620 0 894/1988 0 2439/3219 0
11:31:55 N/A N/A 333/1620 0 857/1988 0 2439/3219 0
11:32:00 N/A N/A 333/1620 0 867/1988 0 2439/3219 0
11:32:05 N/A N/A 333/1620 0 872/1988 0 2439/3219 0

#

Thanks,

FB

[Mel Burslan]

Was this output from top taken while one or more users were launching their initial sessions or while the system was in stable operation ?

Also, is the system continually acting slow or is it just when the users are starting their sessions and launching their applications the first time ?

[Rick Beldin]

I would check for network-related errors. These errors typically cause errors similar to:

XIO: fatal IO error 232 (Connection reset by peer) on X server "fbariselle-m1:0.0"

after 1316 requests (1259 known processed) with 0 events remaining.

XIO: fatal IO error 232 (Connection reset by peer) on X server "fbariselle-m1:0.0"

after 1097 requests (1040 known processed) with 0 events remaining.

XIO: fatal IO error 232 (Connection reset by peer) on X server "fbariselle-m1:0.0"

after 1003 requests (946 known processed) with 0 events remaining.

XIO: fatal IO error 232 (Connection reset by peer) on X server "fbariselle-m1:0.0"

after 1003 requests (946 known processed) with 0 events remaining.

XIO: fatal IO error 232 (Connection reset by peer) on X server "fbariselle-m1:0.0"

after 1002 requests (945 known processed) with 0 events remaining.


This is generally a problem with some sort of network problem. The "connection reset by peer" error is being reported at the socket layer to Xlib. Once Xlib detects a fatal error condition, it terminates the connection.

landiag, network traces, ect. These are the areas where I would focus activities.