netstat error

[Jeremy_7]

I am getting an odd error with netstat...
output is ???? out of memory (file table)
nfile is fine usage is well under set amount
What can I do to make nestat work..
By the way lanscan is hosed too.


-JERM

[Varghese Mathew]

Hi,

You can check the number of files open by using the command # sar -v 3 3 --This command will tell you how many inpde/ no. of files open etc. - so you can have more details on how many files are being used at the moment. In this command output see what is the content below file-sz ov if the number below this "ov" is anything other than "0" then you may need to do some workaround on "nfile" parameter in kernel.

You can even check in the "/var/adm/syslog/syslog.log file - also should have the same errors popped in -- what you are saying . Then you can come to more conclusions..

Hope this helps..
Cheers !!!
Mathew

[someone_4]

Alright .. here we go man .. found some docs lets see if something helps .. what error do you get from lanscan ? Can you copy past the entire out out from both commands? and are there any other commands you cant run ?
########

Problem Description

On two of my servers when I run the netstat -r command, I
get garbage output. It produces hex (hexidecimal) data, and
when doing netstat -n, I get a file table read error.

I have regened the kernel, and the problem remains.

What else should I do to fix this?

Configuration Info

Operating System -HP-UX
Version -9.04
Hardware System -887

Solution

The problem is a mismatch between the kernel that is in
memory and the kernel file itself. Rebooting the system
should fix the problem.
#################
Problem Description

I am running HP-UX 10.20. I get the following error when I attempt to
run the 'netstat' command:

nfile or file not in namelist

Why am I getting this error?

Configuration Info

Operating System - HP-UX
Version - 10.20
Hardware System - HP9000
Series - K460
Subsystem - Networking

Solution

You are getting the error because the '/var/adm/netstat_data' file is
zero bytes.

Perform the following steps to resolve your error:

1. Remove '/var/adm/netstat_data'.

2. Re-execute the 'netstat' command


#########

[Jeremy_7]

Matthew, sar output for ov is 0.
Richard, lanscan output is hwift: bad read
dmesg output Magic number wrong (namelist mismatch?)
netstat -m is
Memory in use:
Mbufs in use:
2164260863 mbufs allocated to
4294904320 mbufs allocated to
6 mbufs allocated to
65536 mbufs allocated to
520380416 mbufs allocated to
1607467008 mbufs allocated to
2067398656 mbufs allocated to
271 mbufs allocated to
16777215 mbufs allocated to
4278846464 mbufs allocated to
16777215 mbufs allocated to
4278846480 mbufs allocated to
1 mbufs allocated to
16777215 mbufs allocated to
4278848519 mbufs allocated to

0 mapped pages in use
0 total bytes of mapped pages in use
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines


rm netstat file ...no louck
reboot no go
running 10.2 on a 9000/K410
also running oracle 7.3.4

[someone_4]

Here we go again ..
Kernal parms for Oracle is what you might need
here is a link with details ..


http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0x00a8663ce855d511abcd0090277a778c,00.html

[someone_4]

Forgot one that might help

http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0xaee3f9beca68d511abcd0090277a778c,00.html

if this is not enough .. post up asking someone on how to set your params for your app.

Richard =)

[ITS-HPUX]

Thank you very much, the Remove '/var/adm/netstat_data' worked.

Thanks

MBrito