I encountered serious problem about OpenVMS

[kokit]

My company use VMS version 6.2, on Feb 9, 2007 we encountered the problem that the system cannot generate any report out.
The error message show as below:
-------------------------------------------
%RMS-F-CHK, bucket format check failed for VBN = 7635905
%TRACE-F-TRACEBACK, symbilic stack dump follows
-------------------------------------------
The attachment contains the screen that I captured from VMS server.

Anyone can help me or know that what is the root cause of this problem?.
Thank you very much for your kindly help or suggestions.

[Uwe Zessin]

Looks like a corruption in an indexed-sequential file. Maybe an RMS expert can help to make the file accessible again.

[Hein van den Heuvel]

The immediate cause for this problem is a corrupted rms indexed file. That file should be restored from backup. Which file is not yet clear. It is likely to be a good sized file (larger than 7635905 blocks).
If you re-run the application you may be able to see it easily with show dev /files (looking for the PID of the re-run)
Or use ANALYZE/SYSTEM ... SET PROC ... SHOW PROC/CHAN

The root cause of the problem is most likely to be hardware related: disk error, power failure, system crash. Check the error logs! There is a very small chance it is a software bug, you are running an older version, but it is not know to have rms indexed file fall over (hmm, except for the CONV/RECLAIM bug that version had).

Restoring a backup is the right approach, but often cause problems missing the most recent data.

Fixing the file MAY be an alternative.
If you are (very) lucky the corruption is not in a main data bucket and a simple CONVERT will fix it. Or maybe a convert will get a good chunk back, and a program can get the data from following the corruption (read by key in Cobol/Basic/DCL)

Anything beyond that requires advance skills. Ask you support organization(s) (HP, Promis) or look around for a company that can help you... if the data is very valuable.

Many other topics deal with this for further reading:
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1071857
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=886747

Just Google for: hein rms chk

But more specifically check out slides 39-50 in:
http://h71000.www7.hp.com/freeware/freeware60/rms_tools/rms_tuning.ppt

Hope this helps,
Hein van den Heuvel
HvdH Performance Consulting
(specializing in RMS indexed file tuning and repair :-)

[Wim Van den Wyngaert]

I have seen disk corruption getting solved by power cycling the disks.

Wim

[Robert Gezelter]

kokit,

As Hein has mentioned, some details need to be checked out.

The first step that I would take would be to use BACKUP to make a backup copy of the file, errors and all.

Then I would, as a precaution, do a full check of the disk with ANALYZE/DISK/READCHECK (the purpose being to identify if there are any physical disk errors).

Then I would, as Hein mentions, determine the extent of the error. It may be nothing more than an alternate index, in which case recovery is (relatively) straightforward, or it could be more extensive, in which case more extensive measures would be needed.

It is hard to give a cookbook answer in the forum, as the correct answer depends upon many details and a careful investigation. Missing a detail could result in more loss of data than is actually necessitated by the damage.

I hope that the above is helpful.

- Bob Gezelter, http://www.rlgsc.com

[Hein van den Heuvel]

Any updates on this problem?
How did you resolve it?
Regards,
Hein.

[kokit]

Thank you everyone very much for the suggestions.

I already tried to do the simple convert data but it cannot success because the system show another error, the error message show as below:
-------------------------------------------
%SORT-F-READERR, error reading DRA0:[PROM55_DBASE]HIST.ISM;1
-RMS-F-CHK, bucket format check failed for VBN=7635905
-------------------------------------------
I also tried to generate the report again for the date 10-11 Feb is OK but for the date 8-9 Feb is not OK, so I think the problem is the data on the date 8-9 is corrupted.

Right now, the most serious problem for me is the space is going up everyday because we cannot do the simple convert data (Compress data).

Could anyone suggest me for this problem?

[Hein van den Heuvel]

>> the system show another error, the error message show as below:
-------------------------------------------
%SORT-F-READERR, error reading DRA0:[PROM55_DBASE]HIST.ISM;1
-RMS-F-CHK, bucket format check failed for VBN=7635905
-------------------------------------------
That really is the very same underlying error.
(btw... for a simple convert you should NOT have needed SORT, just $CONV/NOSORT but that's probably not important just now).

>> I also tried to generate the report again for the date 10-11 Feb is OK but for the date 8-9 Feb is not OK, so I think the problem is the data on the date 8-9 is corrupted.

So write as program/tool to extract the data up to 2/8 and from 2/10 and then worry about 2/8 and 2/9 when you have time/resources.
Use Datatrieve, or Cobol, or DCL, or maybe us a promiss extract procedure?

>> Right now, the most serious problem for me is the space is going up everyday because we cannot do the simple convert data (Compress data).

That doesn't sound to bad, but I would be worried about creating further corruptions.

>> Could anyone suggest me for this problem?
As replied above and in realier replies. If this is beyond the capabilities of your organization then contact a support or consulting organization (such as myself :-)

Regards,
heinvandenheuvel
gmail.com

[Wim Van den Wyngaert]

It could still be a hardware error.
Is your error counter going up (sho er) ?

If so, I would still try a power cycle and hope it is gone. Had already 2 cases were problems were solved this way. 1 on a AS500 and 1 on a 4100.

Wim

[kokit]

I had done total shutdown with power off, Hard Disk and Memory analysis but did not find any error at all; currently we still receive the same error below and in the attached file.

%SORT-F-READERR, error reading DRA0:[PROM55_DBASE]HIST.ISM;1 -RMS-F_CHK, bucket format check failed for VBN=7635905

If you have any other suggestion or recommendation please let me know, thank you very much for you help.