Veritas-bu

[Veritas-bu] socket 24 errors

2001-06-05 12:28:57
Subject: [Veritas-bu] socket 24 errors
From: sixbury AT celeritas DOT com (Sixbury, Dan)
Date: Tue, 5 Jun 2001 11:28:57 -0500
In my case I have the latest Solaris 8 patches, and I actually put the
client_read_timeout to be 7200.  It was 3600.  The only other thing I wonder
about is when the system was rebuilt, (long story short) I had to reinstall
the veritas netbackup software on the solaris server.  I also installed the
latest Netbackup patches as well.  I will look into the duplex on the
system...

Dan

-----Original Message-----
From: Steve Dvorak [mailto:sdvorak AT veritas DOT com]
Sent: Tuesday, June 05, 2001 11:29 AM
To: 'Sixbury, Dan'; EBU (E-mail)
Subject: RE: [Veritas-bu] socket 24 errors


Dan,
The hosts not equal error is not a big deal, NBU works through a process of
reverse look-ups and hostname vs FQDN comparison.  It wouldn't even start
any data transfer if this were a problem.  
One thing I would look at, is if you have an abnormal problem on your
network.  This could be 1/2 duplex vs. full duplex issues, and route issues.
I had a customer with similar types of issues, and after much deliberation
finally replaced the network switch and the problem went away (bad switch).
Also, make sure you are on the latest jumbo patch of Solaris 8.  My
coleagues might laugh at this step, but also try uping the
client_read_timeout=300 to 1200 (20 minutes) in bp.conf.
Steve

-----Original Message-----
From: Sixbury, Dan [mailto:sixbury AT celeritas DOT com]
Sent: Tuesday, June 05, 2001 8:14 AM
To: EBU (E-mail)
Subject: [Veritas-bu] socket 24 errors


ALL:

I am still having problems with our sun server which worked fine when it was
a 2.6, but doesn't work now that it is solaris 8.  The message I always get
is socket 24.  I also noticed that it seems that the backup always fails at
the 400,640KB I even tried excluding the last file it was stuck on and it
still died at the 400,640KB of the backup with status 24.

I changed the backup to backup /opt and it worked fine, but when changing
the backup back to / (root) it would fail.  I included some of the error
messages, and the one that I think that is really odd was the host_equal:
hosts do not match.  Especially since one of the servers is the master and
the other is the media server.  

I can ping each server, I can use bpclntcmd between each server, ftp between
servers, etc. etc.... Communication between the two servers is working.  But
NetBackup isn't.

TIA
Dan


09:01:05 [31206] <16> bpbrm main: from client wildcats: ERR - Cannot write
to STDOUT. Errno = 32: Broken pipe
09:01:06 [31206] <4> bpbrm main: client wildcats EXIT STATUS = 24: socket
write failed

I also thought the log was interesting about hostnames do not match.  But of
course, one is the master server and one is the media server.  In this case
the media server wildcats is backing up to the master server chiefs.

08:46:16 [31206] <4> bpbrm main: INITIATING: version NetBackup 3.4.1GA
200001051
8
08:46:16 [31206] <2> logparams: -backup -mt 2 -to 0 -S chiefs -c wildcats
-hostn
ame wildcats -ru root -cl Unix -sched Full_Backup -bt 991748774 -dt 0 -st 0
-sec
ure 1 -rl 0 -rp 604800 -b wildcats_0991748774 -mediasvr chiefs.celeritas.net
-jo
bid 259 -jobgrpid 0 -masterversion 340000 -maxfrag 0 -v -D 13 -rt 8 -rn 0
-stuni
t TLD891 -cj 1 -pool Unix -kl 28 -rg daemon -ct 0 -use_otm
08:46:16 [31206] <2> LOCAL CLASS_ATT_DEFS: Product ID = 6
08:46:16 [31206] <2> check_remote_allowed: Remote servers allowed
08:46:16 [31206] <2> bpbrm main: send_3_0_client_params = 1
08:46:16 [31206] <2> hosts_equal: hostnames DO NOT compare
08:46:16 [31206] <2> hosts_equal: name1=wildcats  name2=chiefs.celeritas.net
08:46:16 [31206] <2> hosts_equal: hostbuf=wildcats
hostentry->h_name=chiefs.cel
eritas.net
08:46:16 [31206] <2> hosts_equal: compare wildcats with alias chiefs
08:46:16 [31206] <2> hosts_equal: compare wildcats with alias
chiefs.celeritas
08:46:16 [31206] <2> hosts_equal: compare chiefs.celeritas.net with alias
wildca
ts.celeritas.com
08:46:16 [31206] <2> hosts_equal: addresses DO NOT match
08:46:16 [31206] <4> hosts_equal: wildcats=0x32a8a8c0
chiefs.celeritas.net=0x37a
8a8c0
08:46:16 [31206] <2> bpbrm start_bpcd: getservbyname returned port 13782
08:46:16 [31206] <2> getsockconnected: Connect to wildcats on port 740
08:46:17 [31206] <2> bpbrm start_bpcd: DATA_SOCK from bpcr = 4
08:46:17 [31206] <2> bpbrm start_bpcd: NAME_SOCK from bpcr = 6
08:46:17 [31206] <2> bpbrm main: received bpcd success message
08:46:17 [31206] <2> bpbrm main: from client wildcats: read client start
message
08:46:17 [31206] <2> bpbrm spawn_child: /usr/openv/netbackup/bin/bptm bptm
-w -c
 wildcats -den 13 -rt 8 -rn 0 -stunit TLD891 -cl Unix -bt 991748774 -b
wildcats_
0991748774 -st 0 -cj 1 -p Unix -ru root -rclnt wildcats -rclnthostname
wildcats
-rl 0 -rp 604800 -sl Full_Backup -ct 0 -v -mediasvr chiefs.celeritas.net
-jobid
259 -masterversion 340000
08:46:17 [31206] <2> bpbrm write_continue_backup: wrote CONTINUE BACKUP on
COMM_
SOCK
08:46:17 [31206] <2> bpbrm main: closing COMM_SOCK
08:49:21 [31206] <2> bpbrm mm_sig: received ready signal from media manager
08:49:21 [31206] <2> bpbrm readline: retrying partial read from fgets ::
09:01:05 [31206] <16> bpbrm main: from client wildcats: ERR - Cannot write
to ST
DOUT. Errno = 32: Broken pipe
09:01:06 [31206] <4> bpbrm main: client wildcats EXIT STATUS = 24: socket
write
failed

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

<Prev in Thread] Current Thread [Next in Thread>