we got the below error today and was not able to start the sybase back up server.

Logging Backup Server messages in file '/sybase15/ASE-15_0/install/TEST_SYBASE_BS.log'
Open Server Error: 16240.10.0: Net-Library routine sybnet_listen() failed in srv__start_net
Network error: status = 23 - Net-Lib protocol driver call to register a listener failed
.
Open Server Server Fatal Error: 16029.20.0: Failed to start any network listeners
.
No server log file open; Using stderr for log.
Apr 21 13:01:26 2010: Backup Server: 1.29.2.1: Unable to start the Backup Server. See Backup Server error log for additional inform
ation.

Resolution

I found that there are already a couple of processes running from the user sybase15 , killed them and restarted the backup server.
worked fine later.

$ps -ef|grep ^sybase15
sybase15 11985 1 0 12:50:45 pts/1 0:00 sh -c /sybase15/ASE-15_0/bin/backupserver -e/sybase15/ASE-1
sybase15 25979 25977 0 12:00:16 pts/3 0:00 -ksh
sybase15 12925 12923 0 12:57:07 pts/4 0:00 -ksh
sybase15 11986 11985 0 12:50:45 pts/1 0:00 /sybase15/ASE-15_0/bin/backupserver -e/sybase15/ASE-15_0/in
$ptree 11985
11985 sh -c /sybase15/ASE-15_0/bin/backupserver -e/sybase15/ASE-1
11986 /sybase15/ASE-15_0/bin/backupserver -e/sybase15/ASE-15_0/in
$kill 11985 11986
$startserver -f RUN_TEST_SYBASE_BS


Hope this helps

Propably the running process , was sybmultbuf ([URL= The port that used running the Backup Server was reserved by sybmultbuf... so you could not restart it.
I was wondering, in case you kill the process (from unix) could have any impact to BS ;
There is any light way to kill the "zombie" process;

Thanks,Manos.

They are not sybmultbuf.They are the backupserver execs.
It is possible that these backupserver processes were using the ports.

I don't forsee any issue in killing the processes except in some cases the shared memory segments may not get removed.

Not sure about the more graceful way of killing a zombie sybase process dataserver or backupserver at operating system level .

Any Ideas,Guys?

Joshi,

As per errorlog posted , I guess backup server was already started and you tried to start it again.

See the

"Server Error: 16240.10.0: Net-Library routine sybnet_listen() failed in srv__start_net"

It was unable to start the listeners, as it was already used by backup server.

You killed the server process, then it freed the port and in second chance you started the backup server.

you can check used port at os level(in SUN Solaris) ,with netstat -a command.


I think , we can kill the process with chance for clanup of resources with kill -15 which send TERM Signal.

Last option is Kill -9 with kill signal forcefull kill.

If we kill the backup server and dataserver process at OS level forcefully with -9, in some cases it creates problem related to shared memory.

I already posted one thread for shared memory issue. Thanks.