cannot telnet to server

[tom quach_1]

Dear all,

we have experienced some problem this Morning.
users complaned that they could not connect to the server and even myself i could not connect for about 10 mins.
when i was able to connect, ran glance and it shown me
the history.
--NETWORK BOTTLENECK
--DISK BOTTLENECK


the syslog.log

Jun 13 08:06:10 sunuxap2 inetd[956]: telnet/tcp: accept: Bad file number
Jun 13 08:06:19 sunuxap2 inetd[956]: telnet/tcp: accept: Bad file number
Jun 13 08:06:29 sunuxap2 above message repeats 2 times
Jun 13 08:07:02 sunuxap2 inetd[956]: telnet/tcp: accept: Bad file number
Jun 13 08:09:02 sunuxap2 telnetd[15164]: getpid: peer died: Error 0
Jun 13 08:07:11 sunuxap2 inetd[956]: telnet/tcp: accept: Bad file number

ran lanadmin -g 0
and it came back with
Inbound Errors=0
Outbound Errors=0

the logfile /var/adm/nettl.LOG00
shown some errors but 2 days ago, so this one would not count.

<2004> 1000Base-T in path 0/0/10/0/0
Detected a faulty or disconnected cable.

**********************Gigabit Ethernet LAN/9000 Networking******************@#%
Timestamp : Mon Jun 11 PDT 2007 13:09:35.770698
Process ID : [ICS] Subsystem : IGELAN
User ID ( UID ) : -1 Log Class : ERROR
Device ID : 2 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

<2004> 1000Base-T in path 0/0/10/0/0
Detected a faulty or disconnected cable.

**********************Gigabit Ethernet LAN/9000 Networking******************@#%
Timestamp : Mon Jun 11 PDT 2007 13:11:23.770696
Process ID : [ICS] Subsystem : IGELAN
User ID ( UID ) : -1 Log Class : ERROR
Device ID : 2 Path ID : 0
Connection ID : 0 Log Instance : 0
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

<2004> 1000Base-T in path 0/0/10/0/0
Detected a faulty or disconnected cable.
sunuxap1:[root]:/var/adm

I've open a case with HP support but could not get much advices.

PLease give me some advices on how trobleshoot this issue or what to look for te next time it happens again.

Thanks in advance!
Tom

[B. Hulst]

Hi,

Check for the number of open files in the kernel and the amount the system is actually using. With glance for example.

Also check if there are not too many connections still left open. With "netstat -a".

Regards,
Bob

[tom quach_1]

Hi Bob,

in glance where could i see the open files.

Check for the number of open files in the kernel and the amount the system is actually using. With glance for example.

maxfiles 2048 - 2048
maxfiles_lim 4096 Y 4096




Also check if there are not too many connections still left open. With "netstat -a".
would you be more specific. i did a line count and it return with 1099.
#netstat -a |wc
1099 6823 94270

what am i looking in netstat -a

Please advice!
Regards,
Tom

[Noble Sebastian]

Hi Tom ,

You can check number of open files

sar -v 1 1

pls check how many lan card is there ?
Which is activated ...check gateway ..?

Now it is working ...?

[tom quach_1]

Thanks Noble

the parameter : nfile

...but would it have anything to do with
--NETWORK BOTTLENECK


Regards,
Tom

[Noble Sebastian]

Hi tom ,
If something was wrong with inetd -

You could try re-creating your inetd.conf file:

mv /etc/inetd.conf /etc/inetd.bak

cp - p /usr/newconfig/etc/inetd.conf /etc/inetd.conf

Modify any settings particular to your environment.....?


regards
Noble Sebastian ...

[tsf_1]

What is the exact error when you try to telnet?..as the error will give you better hint.
Check if the telnet daemon is listening; below show its listening:
# netstat -a |grep -i telnet
tcp 0 0 *.telnet *.* LISTEN

[Torsten.]

How about this message:

<2004> 1000Base-T in path 0/0/10/0/0
Detected a faulty or disconnected cable

Could be a bad NIC, cable or switch port.

[Tor-Arne Nostdal]

Hi Tom
If you can't find any obvious errors in the server itself, but only have connection problems... it could be a reverse dns lookup problem.

See/Search similar threads on this forum

/Tor-Arne

[tom quach_1]

Dear all,

Thanks for your reply.


i think somehow the telnet connection has something to do with error
--NETWORK BOTTLENECK

would someowe gives me me a hint of what to look for?

the erro
<2004> 1000Base-T in path 0/0/10/0/0
Detected a faulty or disconnected cable.
was happened 2 days before the incident, so i would not count it.
Regards,
Tom

[Mridul Shrivastava]

Hi Tom,

Have you checked the network patches and network interface driver version, are they latest...

Sometime it could be a result of improper handling of driver or patch..

[Safarali]

Hi Tom
restart inetd -c

Regards
Safar

[Safarali]

Hi Tom
restart inetd -c
and not working what is the out put of telnet o (zero)
can you login using telnet zero



Regards
Safar

[tom quach_1]

Thanks all for your help.

the telnet connection is working fine. it just happenned 2 time that users could not connect.
my question is i just want to know the causes
and how to prevent it from happening again.
after i was able to connect to the server. ran glance and saw in the history these messages

--NETWORK BOTTLENECK
--NETWORK BOTTLENECK .

WHAT WOULD BE THE CAUSE OF THIS ALERT.

Thanks in advance.
Tom

[skt_skt]

1.Check the network performance chart/statistics(using perfview) to identify if there is any packet storm happened. Example if the normal trend is 2K and it it increases sundeenly to 60K, 80K or higher is a symptom of a possible storm.

2.Check with your network team if there isa ny error reported on swith ports.

3.When the server was not reachable , was the console connection is also disturbed.? (Are both the server IP and console configured on the same swith).

4. Enable a netwrok logging (using nettl)from UNIX level so that you will be able capture the network related errors.