TDP for Oracle Restore too Slow

[hungng89]

Dear everyone

I am performing TDP for Oracle 5.5.1 restore on Oracle 10 G -2 , Solaris 10 .

The RMAN backup and restore using DISK channel ( local ) is OK.
The TDP Backup speed is FAST - 1 Tape Drive 250 GB/hour ( LTO 5 / GibE)
The BA Client Backup and Restore is FAST - 1 Tape Drive 250 GB/hour.
But TDP Restore is really suck , only 10 GB per hour.

I already read and performed many RMAN Tuning guide for backup ( not found anything for restore ) .

After Analyze TSM Instrument API trace , TDP error , DSM error , TSM Server configuration , RMAN sbtio I found no problem.

I am trying to set debug command to Oracle RMAN and found some thing in the Oracle RMAN restore debug trace log .

DBGRPC:          krmxr: callback returned TRUE, skipping sleep
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, ind=0, sid=313
DBGRPC:          krmxr: sleeping for 10 seconds
DBGRPC:          krmxpoq: xc=4311248208, action="0000018 STARTED40", col_l=17, i

Do anyone has any idea on the above log ? Any advices , document link reference will help me alot.

Thanks and Best Regards

 

[chad_small]

Its always hard to diagnose some performance issues with restores. Do you have the client keeping its mount points? When you FTP from the TSM server to the client what is the speed? What is the sendwait time? Also check out this thread they had similar problems.

http://adsm.org/forum/showthread.php?18222-SLOW-RESTORES-to-SOLARIS-10-Clients

 

[hungng89]

Its always hard to diagnose some performance issues with restores. Do you have the client keeping its mount points? When you FTP from the TSM server to the client what is the speed? What is the sendwait time? Also check out this thread they had similar problems.

http://adsm.org/forum/showthread.php?18222-SLOW-RESTORES-to-SOLARIS-10-Clients

Thanks you chad_small , it not simliar , it exactly the same

Thanks and Best Regards

 

[cheffern]

wow - really not sure...
how many sbtio / tape allocations have you made for this?

 

[hungng89]

Dear Cheffern

I used only single channel / tape drive for backup and restore process