06-10-2012 12:00 PM
Kindly find the bellow error while trying to restore some files (we tried this about 3 times from backup taken from the same server and from another server) so please help ASAP:
6/10/2012 8:37:12 PM - begin Restore
6/10/2012 8:37:22 PM - 1 images required
6/10/2012 8:37:22 PM - media 041AIA required
6/10/2012 8:37:26 PM - restoring image dbprod_1339183543
6/10/2012 8:37:28 PM - requesting resource 041AIA
6/10/2012 8:37:29 PM - connecting
6/10/2012 8:37:32 PM - connected; connect time: 00:00:03
6/10/2012 8:37:32 PM - started process bptm (10028)
6/10/2012 8:37:32 PM - mounting 041AIA
6/10/2012 8:37:32 PM - granted resource 041AIA
6/10/2012 8:37:32 PM - granted resource IBMULT3580-TD22
6/10/2012 8:38:01 PM - mounted; mount time: 00:00:29
6/10/2012 8:38:01 PM - positioning 041AIA to file 1
6/10/2012 8:38:01 PM - positioned 041AIA; position time: 00:00:00
6/10/2012 8:38:01 PM - begin reading
6/10/2012 8:42:32 PM - positioning 041AIA to file 2
6/10/2012 8:42:32 PM - positioned 041AIA; position time: 00:00:00
6/10/2012 8:46:24 PM - Error bptm(pid=8960) cannot write data to socket, 10053
6/10/2012 8:46:24 PM - Error bptm(pid=8960) The following files/folders were not restored:
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/applsysd06.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/akx01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/amfx01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/asfd01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/applsysd02.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/arx01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/astx01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/azd01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/ahmd01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) UTF - /u04/oracledb/prdn/prdndata/bicx01.dbf
6/10/2012 8:46:24 PM - Error bptm(pid=8960) more than 10 files were not restored, remaining ones are shown in the progress log.
6/10/2012 9:16:58 PM - restored image dbprod_1339183543 - (socket write failed(24)); restore time 00:39:32
6/10/2012 9:16:58 PM - Warning bprd(pid=8648) Restore must be resumed prior to first image expiration on 9/9/2012 10:25:43 PM
6/10/2012 9:16:58 PM - end Restore; elapsed time: 00:39:46
the restore failed to recover the requested files(5)
Solved! Go to Solution.
06-11-2012 04:48 AM
Your job details show the following two lines:
6/10/2012 8:37:32 PM - connected; connect time: 00:00:03
6/10/2012 8:46:24 PM - Error bptm(pid=8960) cannot write data to socket, 10053
This indicates that it can connect to the client but cannot write data to the specified location - the 10053 network error is just spurious so ignore it - this is a data write issue.
It will be caused by one of 2 things:
1. rights to place files in that location
2. Inability to overwrite the files - as Praveen says they may be live and located database files - so either restore them to an alternate location or, as long as you are totally sure about what you are doing, take down Oracle first to release the lock and restore them - but be very careful - this sounds like a Production database!
Hope this helps
06-10-2012 12:55 PM
06-10-2012 01:09 PM
appreciate your response . from your proposed solution i need really many hours to try solving this issue
06-10-2012 02:11 PM
please advice
donot thing it seems write permission issue. should initiate the restore with the user having admin/db rights
06-10-2012 09:21 PM
A lot of status 24 errors are related to specific OS.
Please share all of the following:
OS on master, media server and client
NBU version on master, media server and client
Type of restore - normal filesystem restore of database files?
Restore to Source client or different client?
Restore to same or different location?
Please ensure all of following log folders exist before trying another restore:
On master: bprd (restart NBU to enable this log)
On media server: bptm and bpbrm
On client: bpcd and tar
Please rename log folders to reflect process name (e.g. bprd.txt) and post as attachments.
06-10-2012 11:27 PM
Os on master is windows 2008,media and all client are aix 6.1
NBU version 6.5.6 for all
type of files db files (oracle)
same location
06-10-2012 11:39 PM
"appreciate your response . from your proposed solution i need really many hours to try solving this issue"
Yes, I am sorry that this is the case.
Almost certainly, Status 24 is not NetBackup, it is a Network problem. As Marianne shoes, many status 24 are related to the operating system (os network settings etc ...) - and as you will see, many of the details I posted are OS related.
You will not get any reasons to the problem in the NBU logs, all you will see is 'cannot write to socket' - NBU does not know what the problem is, as it is out side NBU.
BUT, I have shown two things that can casue the issue, wrong client version and network comms buffer size (but I have never actually seen this casue a status 24 myself)
So, you are really back working through the detiails I posted.
You could wait, someone will come along and suggest do xxx - this might be in the list of details I have posted and you might get lucky and it is the solution, but how long do you want to wait ?
Martin
06-11-2012 12:39 AM
please create a tar directory on client system. /usr/openv/netbackup/logs/
try to restore only 1 .dbf file from requested set of files.
also make sure at client end there is enough of disk space
06-11-2012 12:48 AM
All I am expecting to see in the logs is where the break in communication is - master, media server or client.
If we can pinpoint the break in communication, you can take it up with relevant server owner.
This error looks like comms problem somewhere between media server and client:
Error bptm(pid=8960) cannot write data to socket, 10053
Are you doing db backups using Standard policy, or Oracle policy with RMAN?
One more thing - double-check exact W2008 version. If R2, only supported as from 7,x.
Please schedule upgrade of your environment ASAP - NBU 6.x is nearing EOSL.
06-11-2012 03:53 AM
we are using standard policy but the issue occur in restor! please advice
06-11-2012 04:03 AM
Windows 2008 enterprise Service Pack 2
06-11-2012 04:13 AM
As per my post earlier today:
Please ensure all of following log folders exist before trying another restore:
On master: bprd (restart NBU to enable this log)
On media server: bptm and bpbrm
On client: bpcd and tarPlease rename log folders to reflect process name (e.g. bprd.txt) and post as attachments.
You will need logs to pinpoint break in communication.
At this point it seems like media server -> client comms problem.
In what state is Oracle when doing file-level backups? Hopefully down?
In what state is Oracle when attempting restore to same location? Hopefully down?
06-11-2012 04:33 AM
As the files you are trying to restore are oracle database files to the orginal location, there is a chance Oracle is using the previous version of the existing files.
Try restoring to alternate location once. if it works then you can move the files manually later when oracle is not using these files or when you take the database down.
As you took a file level backup instead of agent backup, netbackup donot have permissions to write dbf files in original location.
--Praveen
06-11-2012 04:48 AM
Your job details show the following two lines:
6/10/2012 8:37:32 PM - connected; connect time: 00:00:03
6/10/2012 8:46:24 PM - Error bptm(pid=8960) cannot write data to socket, 10053
This indicates that it can connect to the client but cannot write data to the specified location - the 10053 network error is just spurious so ignore it - this is a data write issue.
It will be caused by one of 2 things:
1. rights to place files in that location
2. Inability to overwrite the files - as Praveen says they may be live and located database files - so either restore them to an alternate location or, as long as you are totally sure about what you are doing, take down Oracle first to release the lock and restore them - but be very careful - this sounds like a Production database!
Hope this helps
06-11-2012 05:12 AM
appreciate you all ,and will work on all details ASAP