read from input socket failed(636)

hi,

did you make sure that there is a proper IP resolution on those clients after chaning IP address?

if yes.. 

you may need to clean the cache of the clients to make sure they are using the updated IP address

bpclntcmd -clear_host_cache

if this does not work.. what is the version of the master and clients...?

Hello,

thanks for reply,  i already clear host the cashe and i can see the new ip for the master from client side "bpclntcmd -hn master".

NBU 7.5.0.4 for master and clients 7.5.0.4 to 6.5

hi

we may need to start looking on below logs..

: the debug log forbpsynth (on the master server) and for the bptm or the bpdm reader or writer processes (on the media server).

check if you are seeing any errors on the logs... if possible please attach those .  

http://www.symantec.com/business/support/index?page=content&id=HOWTO51179 

hope this helps.. 

Are all backups failing or just one/some?

Please show us all text in Details tab for one of the failed jobs.

Dear All,

the the jobs that failed are Oracle Database and File system database "we don't use a bpsynth".

the strange of this issue, RMAN run and finish the channels, but the main policy wait and when the be run for two hours failed. please see the attached screenshot.

12/15/2012 7:21:55 AM - Info bpbrm(pid=30598) NODE-N is the host to backup data from    
12/15/2012 7:21:55 AM - Info bpbrm(pid=30598) reading file list from client       
12/15/2012 7:21:56 AM - Info bpbrm(pid=30598) starting bphdb on client        
12/15/2012 7:21:56 AM - Info bphdb(pid=30626) Backup started          
12/15/2012 7:21:56 AM - Info bphdb(pid=30626) Processing /usr/openv/netbackup/ext/db_ext/oracle/RMAN/hot_database_backup-dd.sh          
12/15/2012 7:21:56 AM - Info bphdb(pid=30626) Waiting for the child status       
12/15/2012 7:22:36 AM - Info nbjm(pid=17024) starting backup job (jobid=2667716) for client NODE-N, policy DD_Pmedia-lan_ORACLE_NODE-N, schedule Weekly_Full_Online_Sched 
12/15/2012 7:22:36 AM - Info nbjm(pid=17024) requesting MEDIA_SERVER_WITH_ATTRIBUTES resources from RB for backup job (jobid=2667716, request id:{E5F11562-66C8-42A3-8BE0-27A950236A42}) 
12/15/2012 7:22:36 AM - requesting resource NODE-N-hcart3-robot-tld-4
12/15/2012 7:22:36 AM - requesting resource master.NBU_CLIENT.MAXJOBS.NODE-N
12/15/2012 7:22:36 AM - requesting resource master.NBU_POLICY.MAXJOBS.DD_Pmedia-lan_ORACLE_NODE-N
12/15/2012 7:22:36 AM - granted resource master.NBU_CLIENT.MAXJOBS.NODE-N
12/15/2012 7:22:36 AM - granted resource master.NBU_POLICY.MAXJOBS.DD_Pmedia-lan_ORACLE_NODE-N
12/15/2012 7:22:36 AM - granted resource NODE-N-hcart3-robot-tld-4
12/15/2012 7:22:36 AM - estimated 0 Kbytes needed
12/15/2012 7:22:36 AM - Info nbjm(pid=17024) started backup (backupid=NODE-N_1355548956) job for client NODE-N, policy DD_Pmedia-lan_ORACLE_NODE-N, schedule Weekly_Full_Online_Sched on storage unit NODE-N-hcart3-robot-tld-4
12/15/2012 7:22:37 AM - started process bpbrm (30598)
12/15/2012 7:22:38 AM - connecting
12/15/2012 7:22:39 AM - connected; connect time: 00:00:01
12/15/2012 9:00:57 AM - Info bphdb(pid=30626) Script exited with status = 0 <the requested operation was successfully completed>
12/15/2012 9:00:57 AM - Info bpbrm(pid=30598) validating image for client NODE-N       
12/15/2012 9:00:58 AM - Info bphdb(pid=30626) done. status: 0: the requested operation was successfully completed   
read from input socket failed(636)

Sounds like firewall timeout. 

Seems there is more in the environment that has changed than just IP addresses....

Ask your firewall admins to monitor comms while backup is running - they should be able to confirm firewall timeout.

See this post: https://www-secure.symantec.com/connect/forums/sql-backups-run-fine-parent-job-ends-error-636#comment-8085711

The timeout on firewall we set 100 hour until finish testing, also there is many servers take more than 6 hours and success, but those few server still have this issue