cancel
Showing results for 
Search instead for 
Did you mean: 

Oracle archive not getting to ctlfile

DoubleP
Level 5

OS=Linux Master server is veritas 8.2

Recently, Archive backups of a linux client are not making it to the ctlfile step. The "Backup" and "a_arch_D" run successfully, so we do not get any alerts. But since these archives never get to the ctlfile, they are not succesful. The backup selection is "opt/appl/backup/bin/rman/xydb123_arch_disk.sh". 

The DBAs are seeing this when they check their logs:"ORA-27027: sbtremove2 returned error
ORA-19511: Error received from media manager layer, error text:
Failed to remove, xydb123_ARCHIVE_20191127_0800.p1.s52339.1025424034, from image catalog.
ORA-19513: failed to identify sequential file
ORA-27206: requested file not found in media management catalog"

Other archives using the same master and same media server are working fine, only the ones associated with the one client are having this issue. This client was also successful up to fairly recently. No changes to the client that the DBA is aware of.

The DBAs checked the error with Oracle. Oracle support replied:"This issue is coming from Media Management Layer -
RMAN is not able connect to take or to find the file to delete.Please check with Media Management Layer vendor for resolution on this issue.."

Any ideas?

 

1 ACCEPTED SOLUTION

Accepted Solutions

Update: We have those archives working again.

Here's what the solution was:

First, a little background. We use a tan network for most of our NB clients, so that when the backups are running, they won't interfere with the core traffic and vice versa. 

We compared one of the clients that was working with the one that wasn't. On the "bad" client, out DBA changed the REQUIRED_INTERFACE = from the tan version of the client, to the core version, to match the setup of the "good" client. It worked right away.

We still have the mystery of what changed. It had been working for a couple years before it stopped. DBA states that they didn't make any changes; we didn't either.

View solution in original post

8 REPLIES 8

sdo
Moderator
Moderator
Partner    VIP    Certified

Try this first:

Oracle backup failing with status 6, RMAN error shows ORA-27206: requested file not found in media management catalog

https://www.veritas.com/support/en_US/article.100004280

@sdo, thanks for the response.

I just tried the command mentioned in the article. Here's what I got (of course I changed ip and company name):

$ cd /usr/openv/netbackup/bin
$ ./bpclntcmd -pn
expecting response from server master1.company.com
xyzdbbw2t.paychex.com *NULL* 000.00.0.000 12714
$

Null?

sdo
Moderator
Moderator
Partner    VIP    Certified

yeah... hmmm... @DoubleP ... what does this show :

$ ./bpclntcmd -self

@sdo  $ ./bpclntcmd -self
yp_get_default_domain failed: (12) Local domain name not set
NIS does not seem to be running: (1) Request arguments bad
gethostname() returned: xyzdbbw2t.company.com
host xyzdbbw2t.company.com: xyzdbbw2t.company.com at 000.00.0.000
aliases: xyzdbbw2t.company.com 000.00.0.000
$

sdo
Moderator
Moderator
Partner    VIP    Certified

...what that says to me then is that the client is happy about it's own name, but the previous post shows that the Master does not know who the client is.

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

@DoubleP 

You need bprd on the master server and dbclient on the Oracle client to troubleshoot. 

The first error is this: 

Failed to remove, <image>, from image catalog

I found 2 Veritas TNs and 1 Oracle doc (you need registered account to open this one).

One Veritas TN mentions 7.0.1, but the contents is still worth investigating.

https://www.veritas.com/support/en_US/article.100003582

https://www.veritas.com/support/en_US/article.100015746

https://support.oracle.com/knowledge/Oracle%20Database%20Products/2115192_1.html

@Marianne, thanks; I'll check these out (with our DBA). I'll let you and @sdo know what we discover.

Update: We have those archives working again.

Here's what the solution was:

First, a little background. We use a tan network for most of our NB clients, so that when the backups are running, they won't interfere with the core traffic and vice versa. 

We compared one of the clients that was working with the one that wasn't. On the "bad" client, out DBA changed the REQUIRED_INTERFACE = from the tan version of the client, to the core version, to match the setup of the "good" client. It worked right away.

We still have the mystery of what changed. It had been working for a couple years before it stopped. DBA states that they didn't make any changes; we didn't either.