While updating all indexes for a file the process runs ok until at completing the update of a relational index. The workstation shuts arev off and returns to windows.

Upon re-entry to arev, I attempt a query on the window/file that caused the problem and the message below is returned

Message 521: Another user is currently updating indexes on the file

!filename 18 jun 1993 waiting for this lock to be released …

Workstation WIN98, server NT4.0, Arev 2.03

Any help would be appreciated.

Pat at Escrow, Inc.

Patrick - do you have Win 98 write-behind caching enabled? If so, disable it.

To check, right-click My Computer on the Win 98 workstation, then select Properties, and you'll find a Filing System / Performance options tab where you should uncheck the checkbox marked "enable write-behind caching on all drives".

Steve

The Sprezzatura Group

World Leaders in all things RevSoft

Congratulations to Sprezzatura for having the first post of the new year as indicated by the webserver internal clock. Sprezzatura unseats mr. Barry Stevens, last year's first poster.

A heartfelt congratulations to all at Sprezzatura!

Happy New Year!

Congratulations to Sprezzatura for having the first post of the new year as indicated by the webserver internal clock. Sprezzatura unseats mr. Barry Stevens, last year's first poster.

Figures…both year's winners were Aussies. Unfair advantage IMHO

[email protected]

SRP Computer Solutions, Inc.

Ahh but it's the webserver's clock - you could always post at 00:01 ….

The Sprezzatura Group

World Leaders in all things RevSoft

All I can say is…

It pays not to host while pung over….

Was I the last one to post in 2002?

The setting in win 98 was already correct.

Additional info: I have a btree and relation on the same field that is causing the problem. The month end balancing report depends on the relational index.

Pat

Question, what should acompany the original one (not the year strat/end one:) )

How can be indexing changed that it will report the station locking records … the unwanted indexor.

Or in general how can I obtain ID of locking station of an arbitrary record.

(I expect it is not too hard to found answers in the discussion)

:) Even I am not able to clasify this

… is it an answer or a question :)

"Or in general how can I obtain ID of locking station of an arbitrary record?"

Can be done in Novell using a few arcane practices (bindery support and enduser console operator status and IPX and Logical locking driver). This is a little complicated - See http://www.revelation.com/__8525652b0066bfaf.nsf/0/e6a3fad0d8b2c7a9852566c400312d4a?OpenDocument

This functionality is *NOT* supported by the Windows API in any sense, so it either must be devised by Revelation Software (say for the NLM / LH Windows 2000 service diagnostic modes) or remain unsupported.

There are some diagnostic modes for these drivers which may assist, or else you could write a MFS to intercept lock/unlock calls and represent them somewhere programmatically yourself.

The Sprezzatura Group

World Leaders in all things RevSoft

This is the only workstation accessing arev all other workstations have been loged offed. Therefore tring to determine which workstation is causing the problem. I do not think is the problem or am I wrong in my assumption?

Well the only other process that might lock these files would be a virus scanner (?)

What network driver are you using? (Run WHO at TCL to display)

The Sprezzatura Group

World Leaders in all things RevSoft

Another thougut is a Backup Program (maybe running from a user-workstation) or on the Server. I've seen older versions of Backup Exec that lock files during backup. This caused problems with dedicated indexers.

Don M.

I am not an expert in locking …

It seems to me that locks on Novel remain for a while even after logging off.

Logging on the workstation as another user releases the locks.

Try the Aaron code WHOSELOCK('!File','') … I hope it will identify the locker.

Thanks for all the help but I think we are going in the wrong direction. Let me reexplain the problem.

During the process of rebuilding all indexed fileds for a file the btrees rebuild ok but during the rebuild of a relational index the rebuild gets to 99% at which time the system crashes NOT MY APPLICATION BUT REVELATION the workstation returns to the desktop.

The only workstation that was loged-in was the workstation that was used to perform the index rebuild.

Queries performed on this file before the rebuild process work fine.

upon reentring revelations on the same workstation an attempting a query on the file I get the message "another user currently updating indexes.

I already know which workstation created the lock what I do not know is WHY the system crashes (abnormal termanation of the session) during the rebuild index process.

Sorry, but maybe I did not explain the proble correctly the first time

Thanks again for all the help.

OK, now I understand the question/problem.

(The lock issue was only the result of previous problem.)

Sorry, I cannot help you as I have no experience with relational indexes.

Try this from the workstation that aborted:

F5 (TCL)

Type: EVAL UNLOCK ALL and press the key.

That should clear the workstation's lock

Don M.

Don,

It seems to me, that the Patrick's problem is not with unlocking.

The problem is that indexor crashes during relational index update.

Hippo you are correct the cause of the problem is in the updating.

Don your ideal does help to remove the effect of the problem. Thanks

Pat

View this thread on the forum...