Server session getting disconnected

[marvik]

Hi Friends,

I am trying to do some file transfer across servers and its getting hanged

My putty connection is also getting disconnected intermittently

ServerB
OS :: 11.11
Patch Level ::
BUNDLE B.11.11 Patch Bundle
BUNDLE11i B.11.11.0102.2 Required Patch Bundle for HP-UX 11i, February 2001

OnlineDiag B.11.11.04.09 HPUX 11.11 Support Tools Bundle, Sep 2001
HWEnable11i B.11.11.0109.6 Hardware Enablement Patches for HP-UX 11i, September 2001

Detail::
1. when I scp files from ServerA to ServerB
Example::

ServerA:\> scp HP-UX_B.11.11_Secsh.depot 160.110.231.45:~vkarmak/
Couldn't connect to PRNGD socket "/var/run/egd-pool": No such file or directory
HP-UX_B.11.11_Secsh.depot 1% 128KB 23.7KB/s - stalled -Killed by signal 2.

in syslog I get some errors like ::
sshd[131]: error: accept: No buffer space available
inetd[661]: ftp/tcp: accept: No buffer space available
inetd[661]: chargen/tcp: accept: No buffer space available
syslog: bad auth protocol start: EOF
inetd[661]: discard/tcp: accept: No buffer space available
inetd[661]: login/tcp: accept: No buffer space available
inetd[661]: printer/tcp: accept: No buffer space available
telnet[10283]: getpid: peer died: Error 0
inetd[661]: ident/tcp: Died on signal 13
inetd[661]: dtspc/tcp: accept: No buffer space available
inetd[661]: exec/tcp: accept: No buffer space available
inetd[661]: klogin/tcp: accept: No buffer space available
inetd[661]: shell/tcp: accept: No buffer space available

I have also rebotted my server but no good.Below is ulimit -a o/put incase it may help u
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 1048576
stack(kbytes) 131072
memory(kbytes) unlimited
coredump(blocks) 4194303
nofiles(descriptors) 2048

Thanks for your help

Marvik

[Shrikant Lavhate]

Hi,

Try to split file you want to transfer to smaller files and then transfer and at the destination again join all files to create original one.
You can use split command for that. For more details man split.

-=ShRi=-

[marvik]

Thanks Shrikant,

But the problem is not related to split i believe, coz 150KB or 9MB files isn't a size I need to split for transfer

[Yogeeraj_1]

hi Marvik,
can you try to pull the file from the other side?

e.g. From server B
scp SERVERA:/depot/HP-UX_B.11.11_Secsh.depot /depot/

revert

kind regards
yogeeraj

[Shrikant Lavhate]

Hi,
>My putty connection is also getting disconnected intermittently

Check for any network issue or timeout settings.

>ftp/tcp: accept: No buffer space available

check your memory utilization details.

[melvyn burnard]

I think you have an underlying resource issue:
sshd[131]: error: accept: No buffer space available
inetd[661]: ftp/tcp: accept: No buffer space available
inetd[661]: chargen/tcp: accept: No buffer space available

This is basically teling you that there is a shortage of buffer space for inetd to use, or put another way, you are having memory resource issues.
I suggest you check you system memory uiseage.

[marvik]

Hi,
Thanks for so many responses
I read in forum to tune some ndd parameter but I dont know which.There hasn't been any changes and this problem erupted suddenly.

Here is top output showing free memory ::
Memory: 2661408K (771384K) real, 3450832K (1354420K) virtual, 3779948K free Page# 1/8

do u think I need to tune some kernel parameter or TCP parameter

[marvik]

Hi Yogi,

Yes I tried the other way as well but no good.

[SUDHAKAR_18]

Hello,

Is your CPU and memory utilisation is normal ? why cant u update the bundles to latest to Dec2007.

Keep updating..
Regards,
SHR

[marvik]

Sudhakar,

Its there in the back of my mind but I need some turnaround immediately patching will need download of stuff and then downtime from business

[SUDHAKAR_18]

what about CPU and memory utilisation?

[Yogeeraj_1]

hi again,

Can you try to mount the directory using NFS and cp command to transfer the file?

kind regards
yogeeeraj

[Ian Kidd_1]

the errors in your syslog all seem to be tcp based. could you run a `netstat -s -p tcp` ? I'm curious what your "connect requests dropped due to full queue" looks like.

[marvik]

Thanks Ian ,
Here is the o/put of tcp -s -p tcp

tcp:
38130073 packets sent
37745917 data packets (1004063245 bytes)
129328 data packets (39704809 bytes) retransmitted
383940 ack-only packets (64366 delayed)
10 URG only packets
5 window probe packets
11 window update packets
714553 control packets
20303184 packets received
18707306 acks (for 1003928592 bytes)
626882 duplicate acks
0 acks for unsent data
1186178 packets (87837205 bytes) received in-sequence
9 completely duplicate packets (9562 bytes)
374 packets with some dup, data (239628 bytes duped)
5585 out of order packets (1436135 bytes)
2 packets (4225131854 bytes) of data after window
292 window probes
803427 window update packets
138 packets received after close
0 segments discarded for bad checksum
0 bad TCP segments dropped due to state change
151322 connection requests
156114 connection accepts
307436 connections established (including accepts)
311881 connections closed (including 2063 drops)
534 embryonic connections dropped
18371900 segments updated rtt (of 18371900 attempts)
103805 retransmit timeouts
26 connections dropped by rexmit timeout
5 persist timeouts
3292 keepalive timeouts
3238 keepalive probes sent
13 connections dropped by keepalive
0 connect requests dropped due to full queue >>>>>>>>
564 connect requests dropped due to no listener

[Yogeeraj_1]

hi marvik,

can you post the output of the following command?

glance -t

e.g.
--------------------------------------------------------------------------------
SYSTEM TABLES REPORT Users= 2

System Table Available Used Utilization High(%)
--------------------------------------------------------------------------------
Proc Table (nproc) 4096 175 4 4
File Table (nfile) 63498 648 1 1
Shared Mem Table (shmmni) 512 10 2 2
Message Table (msgmni) 4096 2 0 0
Semaphore Table (semmni) 4096 21 1 1
File Locks (nflocks) 4096 26 1 1
Pseudo Terminals (npty) 60 0 0 0
Buffer Headers (nbuf) na 52938 na na



revert!

kind regards
yogeeraj