03-10-2011 07:22 AM
Hi,
I was restoring our catalog though tape,
during first attempt we got Media Read Error after restoring around 30 GB of data, and the restoration failed.
Again during the second attempt and through different tape, i got the below error
Error bpbrm socket read failed: errno = 52 -Stream ioctl timeout
Error bptm media manager terminated by process
Error bpbrm client restore Exit Status 13: file read failed
I am really not understanding why this happened, if it is media read issue/tape drive issue or any other network issue (network issue might not be there since we were restoring it through SAN tape librarydrives directly connected to it)
Can anyone help me to trace the issue?
We followed the below steps while restoration;
* Taken Full Catalog Backup
* Gave new IP /hostname to new server.
* Installed Master Server and releated softwares and made the netbackup patch level same as of old.
* started recovery thorugh GUI
03-10-2011 08:49 AM
Note: Do not change the host name of a NetBackup server. This practice is not recommended because it can be necessary to import all previously used media to the server before you can use it under the new host name.
Changing or modifying any NetBackup server host names must be done with extreme care, whether they be Master Servers or Media Servers.
It is highly recommended that any NetBackup Server host name changes be done through Symantec's consulting services. This is due to issues such as the server host name being appended to the images in the catalog. Simply changing the Master Server host name can cause restores to fail, connection problems with clients and having to import all media back into the catalog.
Changing or modifying the master host name is not supported by Technical Support without the assistance of a qualified Symantec Consultant on site.
http://www.symantec.com/docs/TECH32599
http://www.symantec.com/docs/TECH31385
03-10-2011 09:23 AM
Please share the following info:
NBU version
OS version
Hot or cold catalog restore? Please share exact process that was used for catalog restore. If hot catalog restore - full or partial restore?
I'm surprised that ANY data was actually restored with different hostname.
As Chris pointed out - hostname change is not supported without Symantec consulting engagement.
03-11-2011 07:06 AM
Sahil,
dont change hostname
03-14-2011 08:58 AM
Very Sorry,, i really appologise for my typing mistake,
We followed the below steps while restoration;
* Taken Full Catalog Backup
* Gave new IP to the new server. Kept the same hostname that of old master server.
* Installed Master Server and releated softwares and made the netbackup patch level same as of old.
* Configured the drives and robots.
* started recovery thorugh GUI.
I did this activity for four times by using different medias as well as different tape drives, but every time i am getting similar errors;
Error bptm (pid=16763) cannot read image from media id , drive index 5, I/O error
Error bpbrm (pid=16751) from client bkpsvr: more than 10 files were not restored, remaining ones are shown in the progress log.
Error bpbrm (pid=16751) client restore EXIT STATUS 85: media read error
I want to understand where exactly the issue would be.
Hi Marriane;
My NBU Version is 6.5.5
OS is HP UX V2 but where we are migrating is HP UX V3
And we are doing FULL HOT catalog restore.
03-14-2011 01:27 PM
Your last post has a different error to your opening post?
You need to enable logs to troubleshoot: bptm and bpbrm
Create these directories under /usr/openv/netbackup/logs
Please also enable Media Manager logging by adding VERBOSE entry to /usr/openv/volmgr/vm.conf.
Restart NBU. Device errors will be logged to /var/adm/syslog/syslog.log.
***************************
Another thought - Device configuration is somewhat different in 11i v3. Unfortunately the 6.5 Device Config Guide does not contain specific info for 11i v3.
I suggest you use instructions in NBU 7 Device Configuration Guide http://www.symantec.com/docs/TECH127069
Also - remember to disable SPC-2 SCSI reserve as well as EMS Tape Device Monitor (see Device Config Guide).
03-15-2011 09:51 PM
Hi,
In 11 i V3, SPC-2 SCSI parameter is disabled by default. I disabled EMS Tape Device monitor, also all the logging level is increased, but still facing the similar issue. After restoraion of around 25 GB catalog, restoration is failing. Also i observed that while restoring image of a particular client the restoration is failing.
Even i observerd below errors in bptm logs;
<4> db_error_add_to_file: VBRC 2 2077
<4> db_error_add_to_file: cannot read image from media id , drive index 8, I/O error
<16> read_data: cannot read image from media id , drive index 8, I/O error
Also i bpbrm logs i find;
<16> bpbrm main: from client : UTF - /netbackup/openv/netbackup/db/images/xxxx/1292000000/catstore/_1282646236_FULL
03-15-2011 09:58 PM
Any device-related errors in /var/adm/syslog/syslog.log?
03-15-2011 10:51 PM
As such didn't see any any error in syslog.Not sure if the line made bold points to any error;
Below is syslog.log,
tldd[3288]: TLD(0) MountTape XXXX on drive 8, from slot 87
tldcd[3314]: returned from mm_authenticate_request, vauth_action=4, VAUTH_DENIED=1, VAUTH_ALLOWED=0
tldcd[3314]: tldcd.c.3081, process_request(), received command=1, from peername=bkpsvr, version 50
tldcd[3314]: Processing MOUNT, TLD(0) drive 8, slot 87, barcode xxxxxL3 , vsn XXXX
tldcd[4426]: TLD(0) opening robotic path /dev/rchgr/autoch1
tldcd[4426]: inquiry() function processing library HP ESL E-Series 7.00:
tldcd[4426]: TLD(0) initiating MOVE_MEDIUM from addr 12374 to addr 4103
tldcd[4426]: TLD(0) closing/unlocking robotic path
tldcd[3314]: inquiry() function processing library HP ESL E-Series 7.00:
tldcd[3314]: tldcd.c.2695, newfd = INVALID_SOCKET, newfd=-1, timersig=1, error=4, EINTR=4, selectret=-1
inetd[4461]: registrar/tcp: Connection from xxxxx (xxx.xx.xx.xx) at Wed Mar 16 09:57:53 2011
tldd[3288]: DecodeMount: TLD(0) drive 8, Actual status: STATUS_SUCCESS
ltid[3261]: LTID - received ROBOT MESSAGE, Type=54, LongParam=0, Param1=8, Param2=0