Restore of archivelog from OSB via zdlra

From: LK <lkaing_at_gmail.com>
Date: Fri, 27 Jul 2018 15:51:10 +1000
Message-ID: <CAPqBzac-fDPez2LBSeawQ53L7Uww4sgPs4P2S3ZyVkzgrT3d7w_at_mail.gmail.com>

Hi all,

RA Library 12.2.0.1.0
obtool version 12.1.0.2.0
ra_automation-12.1.1.1.8.201803-27981449.x86_64 Protected DB version: 11.2.0.4

I'm taking RMAN backups to zdlra, and copying the backuppieces to tape. The backup and restore to and from ZDLRA works perfectly.

I'm now trying to utilise the copy-to-tape facility using OSB. The tape drive model is HP Ultrium LTO 5. The tape library model is SL30000 managed by ACSLS. The backup, validate and crosscheck of the backuppieces on OSB. But the restore hangs. We were told that the bug should have been fixed in the RA_automation 12.1.1.1.8.201704 but it does not appear to be the case.

I have a few issues that I hope someone can help me with here.

1. I've set retention policy on the RA to one day and 2 weeks on tape (OSB), however, all backups still exist in both the RA and tape. I have not been able to recreate the scenario where backups exist only on tape. I've reduced the space reserved for the protected database and it has not helped. What is the trick to getting the retention policy working as I need here?
2. In order to force a restore of a backuppiece on osb I've had to delete the backuppiece in the RA. A validate works but a restore hangs.

The sbtio.log shows the below repeatedly:

Status => 449
Reason => Retry Request
<?xml version="1.0"?>
<Error>
<Code>Retry Request</Code>
<CodeDescription></CodeDescription>
<Message>ORA-45200: HTTP status code: 449
ORA-45210: resource busy, retry possible </Message>
<Retry>true</Retry>
<Busy>true</Busy>

</Error>

An RMAN debug shows the thing sleeping for 10 seconds and goes in a loop

  4767 DBGRPC:                 2 STEPstepid=2 cmdid=1 status=STARTED

devtype=SBT_TAPE instid=2 parms=SBT_LIBRARY=/u01/app/oracle/admin/ common/zdlra/dev/lib/libra_2018.so,ENV=(RA_WALLET='location=file:/u01/app/oracle/admin/common/zdlra/dev/wallet

credential_alias=        zdlraingest-scan:1521/zdlra:dedicated')
   4768 DBGRPC:           krmqgns: no work found for channel ORA_DISK_1
(krmqgns)
   4769 DBGRPC:            (krmqgns)
   4770 DBGRPC:          EXITED krmqgns with status 1
   4771 DBGRPC:          krmxpoq - returning rpc_number: 8 with status:
STARTED148 for channel ORA_SBT_TAPE_1
   4772 DBGRPC:          krmxr - sleeping for 1 seconds
   4773 DBGRPC:          ENTERED krmqgns
   4774 DBGRPC:           krmqgns: looking for work for channel default
(krmqgns)
   4775 DBGRPC:           krmqgns: commands remaining to be executed:
(krmqgns)
   4776 DBGRPC:           CMD type=IRESTORE cmdid=1 status=NOT STARTED
   4777 DBGRPC:                 1 STEPstepid=1 cmdid=1 status=NOT STARTED
devtype=SBT_TAPE bs.stamp=981899610 step_size=27821056 Bytes
   4778 DBGRPC:                     1 DON_F stepid=2 satisfied=FALSE
   4779 DBGRPC:                     2 DON_F stepid=3 satisfied=TRUE
   4780 DBGRPC:                 2 STEPstepid=2 cmdid=1 status=STARTED

devtype=SBT_TAPE instid=2 parms=SBT_LIBRARY=/u01/app/oracle/admin/ common/zdlra/dev/lib/libra_2018.so,ENV=(RA_WALLET='location=file:/u01/app/oracle/admin/common/zdlra/dev/wallet credential_alias= zdlraingest-scan:1521/zdlra:dedicated')

I do not see any restore jobs scheduled in osb to even attempt a restore.

I have only ever successfully performed 1 single archivelog restore in all of my tests. And cannot seem to get it to restore ever again. It simply hangs. I've waited for 20 minutes before killing it.

Are there any parameters that you need for this to work all the time.

Is there anybody using ZDLRA and OSB out there? Are you able to restore successfully?

Thank you,

Leng.

--
http://www.freelists.org/webpage/oracle-l

Received on Fri Jul 27 2018 - 07:51:10 CEST