listener hung

[Steve Bear_1]

$cat listener.ora
LISTENER =
(ADDRESS_LIST =
(ADDRESS=
(PROTOCOL=IPC)
(KEY= db01)
)
(ADDRESS =
(PROTOCOL = TCP)
(HOST = db01)
(PORT = 1526)
)
)
LOG_DIRECTORY_LISTENER = /apps/oracle/product/7.3.2.2/network/log
LOG_FILE_LISTENER = sqlnetv2
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = dw01)
(ORACLE_HOME=/apps/oracle/product/7.3.2.2)
)
)
TRACE_LEVEL_LISTENER = ADMIN

$more tnsnames.ora
live =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS =
(PROTOCOL = TCP)
(HOST = db01)
(PORT = 1526)
)
)
(CONNECT_DATA =
(SID = dw01)
)
)

[Steve Bear_1]

no, just one ORACLE_HOME

[T G Manikandan]

Can you change db01 and give the ip-address instead of hostname.

please check that.

Can you do a

#nslookup db01

REvert

[Steve Bear_1]

did try though, but no luck.

[T G Manikandan]

What is the output of ..
#nslookup db01
#nslookup

[Graham Cameron_1]

Just noticed "Even svrmgrl also gets hung."

At what point?
When you run it up or when you connect internal?

-- Graham

[Yogeeraj_1]

thinking the microsoft way.

did you try to restart the server?

[T G Manikandan]

Any messages from alert.log of Oracle and /var/adm/syslog/syslog.log files.

[Yogeeraj_1]

do we need an "inetd -c" ?

[Francesco_13]

the command "lsnrctl status" run correctly?

[Jean-Luc Oudart]

Hi Steve,

Something completely different.
It happened to us a few weeks ago. One of the internal drive was dying. I mean it was not dead. (nor alive !).

For whatever reason, it did affect our listener, as we were not able to shut it down.

The problem was fixed when internal disk replaced.

Regards,
Jean-Luc

PS : our internal disks are mirrored. But because the disk was not "dead" this created more pb.
Check you syslog.log for any messages.

[Jean-Louis Phelix]

Hi,

What about tusc'ing your listener just to get the step where it stops ?

Regards.

[Yogeeraj_1]

Example:
tusc -f -o /tmp/tusc.log "lsnrctl status"

[Dave Hutton]

Are there more then 1 network cards? Does your netstat -rn look okay? Maybe you have some funky issues with routing?
Your arp table cleared? Maybe your arp is still regestering your old mac/ip addresses?

I'm also running out of ideas.

Dave

[Bryan D. Quinn]

Hey Steve,

This might sound simple, but just to step back for a minute. Can you ping the new ip address from the box itself? If so, can you ping the boxes hostname from the box?

-Bryan

[Steve Bear_1]

Thanks for your feedback i did try and check most of the thing but no luck so far.

I've logged a TAR n waiting for a feedback.

Many thanx,

[Bryan D. Quinn]

So you were able to ping both?

-Bryan

[Steve Bear_1]

Yes.

[Hari Kumar]

i had gone through all of our collegues answers they all shared there wonderful ideas
my perception is that:
Check whether any software or application hit against the port 1521 of listener and if so then change the listener port or application port.
mixing 32 and 64 bit databases and listeners is also one of the cause
check for listener.log file size it should not be more than 2 Gb and trim the log file to as small as possible

[Dave Hutton]

Like Hari suggested, you can use lsof to easily find if something maybe also using that port. Which can be downloaded at:
http://hpux.cs.utah.edu/hppd/hpux/Sysadmin/lsof-4.69/

Dave