VMBXAR0020S Error 0020 occurred opening file CAAFCCFB 78CA5C8S.

Products

VM:Archiver for z/VM VM:Backup for z/VM VM:Backup High-Speed Disaster Recovery Option (HiDRO) Mainframe VM Product Manager VM:Manager Suite for Linux on Mainframe VM:Manager Suite for z/VM VM SUITE

Issue/Introduction

Description:

VM:Backup was attempting to run a VMARCH ARCHIVE job. While processing this job, one of the files on VMARCH's Stage disks gave VM:Backup an error.

Solution:

Due to the nature of this error, every time a VMARCHIVE ARCHIVE job is run, it will continue to contain this error.

Cancel the VMARC job from VMARCH

VMARCH CAN JOB E5DB5087 <=JOBID

Warning: Try to complete the rest of these steps before another VMARCH ARCH command is issued. Instructions that follow do not assume that this has happened.
Run VMARCH CHECK DBDASD.

This reports on files that are on DISK but not in the DATABASE or files that are in the DATABASE but not on the DISK.

In the report you should see the file referenced in the message. For example:

Files referenced in the database but not found on DASD:

Owner Filename Filetype Archid DASDFN DASDFT vaddr
-------- -------- -------- ----------------- -------- -------- -----
VMANAGER CMS EXEC CAAFC932.1 CAAFCCFB 78CA5C8S 01C0
```
Owner          Filename  Filetype  Archid             DASDFN      DASDFT      vaddr--------       --------  --------  ----------------   --------    --------    -----VMANAGER       CMS        EXEC     CAAFC932.1         CAAFCCFB    78CA5C8S    01C0
```
Note: this is the file in the message:
VMBXAR0020S Error 0020 occurred opening file CAAFCCFB 78CA5C8S.
Issue the PURGE command against this file:

VMARCH PURGE F CMS EXEC ( ARCHID CAAFC932.1 NOGRACE

This will purge the file with no grace period. It will be gone from the database. If there were any other files in the CHECK DBDASD report from step 2, run the same PURGE command against them.
You may get an error:

VMAPUR0550E File CMS EXEC, archid CAAFC932.1, is included in a VM:Backup job; it cannot be purged at this time.
Ready(00550); T=0.01/0.01 11:40:15

Verify there are NO outstanding VMARC jobs in VMARCH. If so, cancel that job from the VMARCH side. An outstanding job with this file in it would not allow the PURGE command to complete. (see Step 1)
```
VMARCH QUERY JOBS                                       Submit                    EndType     Jobid          Status         Date        Time          Date        Time-------  --------       ---------      --------------------      --------------------ARCHIVE  E5DB5087       NO_VMBJOB      01/18/2013  12:43:08      --/--/----  --:--:--Ready; T=0.01/0.01 11:37:15
```
NO-VMBJOB VM:Archiver requested information on a job for which VM:Backup has no information.

This status can be caused by either a COLD start of VM:Backup or by changing the VM:Backup configuration information in VM:Archiver while the job is being processed. If the problem was due to a configuration error, event 9005 will attempt to restart processing for this job. If event 9005 is unable to restart processing for the job, issue the VM:Archive CANCEL command to cancel the job.
If you have this situation, you must CANCEL the job and then issue the PURGE command before another ARCH job is submitted.
Another by-product of this situation can be the following message from a VMARCH ARCH job:

10:34:42 VMBTMP0300E Job 62B35C83 has ended abnormally for the following reason:
10:34:42 VMBTMQ0551E Tape cannot be positioned. It contains too few data sets.
10:34:42 VMBJIM0999I Backup job 62B35C83 (request 00000001) has been placed on h
10:34:42 VMBJIN0388I Rolling back checkpoint records for job 62B35C83.

It's the ARCHIVE job that only contains a file that cannot be found. VM:Backup realizes that there is nothing to write so it doesn't create the newdataset on the tape. The Keyword Data Report from that run, does not tell VM:Archiver that a DSN 21 (for example) was not written on the tape. Therefore, VM:Archiver doesn't realize that DSN 21 was not written and should request in the next ARCHIVE job that DSN 21 be created. However, that is not being done, so, the next ARCHIVE job specifically tells VM:Backup to write DSN 22. VM:Backup then issues a message that too few datasets exist and it currently only has 20 DSN's, although VM:Archiver is asking to write #22.

The solution is to cancel the VMARCH job and close the TAPESET using VMARCH command: VMARCH TAPE CLOSE.

After that, you can submit the ARCHIVE job again. This time a scratch tape will be requested.

Environment

Release: VMBKUL55400-3.5-VM:Backup
Component: