Essbase Invalid Block Header Errors 
Author Message
 Essbase Invalid Block Header Errors

Recently we have experience two instances of
Invalid Block Header Errors in Essbase 5.0.1
Patch 7 running on Window NT 4.0.

The most recent problem seems to be related to a
user of our VBA Excel application Cntrl Alt Del
while she was retrieving data from the server.
This caused locks on the database that were un-
removable.  The only way to remove was to go to
the server and kill the application and restart
it.

Has anyone else experience similar findings with
Essbase?

Sent via Deja.com http://www.***.com/
Before you buy.



Mon, 22 Apr 2002 03:00:00 GMT
 Essbase Invalid Block Header Errors

We had a similar Block header error and we were
able to reolve it by installing Patch 10.

Quote:

> Recently we have experience two instances of
> Invalid Block Header Errors in Essbase 5.0.1
> Patch 7 running on Window NT 4.0.

> The most recent problem seems to be related to a
> user of our VBA Excel application Cntrl Alt Del
> while she was retrieving data from the server.
> This caused locks on the database that were un-
> removable.  The only way to remove was to go to
> the server and kill the application and restart
> it.

> Has anyone else experience similar findings with
> Essbase?

> Sent via Deja.com http://www.deja.com/
> Before you buy.

Sent via Deja.com http://www.deja.com/
Before you buy.


Tue, 23 Apr 2002 03:00:00 GMT
 Essbase Invalid Block Header Errors
Most of the time Invalid Block error are not an  Essbase problem if it is a
high patch but an hardware problem.

It is often generate while Essbase is running on Raid server with an old
bios or controller version.

If the Invalid Block is an agregate block , just do a clear upper block and
launch an agregation.

If the invalid block is level zero block u must do a clear all :(. Actually
it is not able in Essbase to destroy a specified block. I hope u did a
Validate before making a backup of the database

                                    Jerome PARRAIN
                                    SIS France



Quote:
> Recently we have experience two instances of
> Invalid Block Header Errors in Essbase 5.0.1
> Patch 7 running on Window NT 4.0.

> The most recent problem seems to be related to a
> user of our VBA Excel application Cntrl Alt Del
> while she was retrieving data from the server.
> This caused locks on the database that were un-
> removable.  The only way to remove was to go to
> the server and kill the application and restart
> it.

> Has anyone else experience similar findings with
> Essbase?

> Sent via Deja.com http://www.deja.com/
> Before you buy.



Tue, 23 Apr 2002 03:00:00 GMT
 Essbase Invalid Block Header Errors
This is the steps that you need to take to fix the Invalid Block Header
(IBH) problems:
Stop the database if it is running.
Delete the <db-name>.IND file.  Make sure that it is NOT ess00001.IND
file
Start the database.  When the database's Essbase Kernel starts, it will
rebuild the Data Freespace Table, eliminating any corruption there.
Run Validate.  If Validate does not generate an IBH, then the database
does not suffer from this corruption.
If Validate generates an IBH, then start over with a new, empty
database.

For more information read the following:

History Originally, this Storage Manager error was thought to be a
problem resulting from a bug in our
code. Through many patches and code fixes, we believe all errors in our
code that could lead to this error
have been resolved . We are now 99% confident that the bug now lies in
the O.S.'s hardware layer in
environments that have Essbase servers writing data to some type of RAID
configuration.

Common to all Experiencing this bug: Customers who run RAID devices. A
Raid (Redundant Array of
Inexpensive Disks) is basically stored external data up and above the
internal amount of slots allocated
inside of a normal server.

Four cases that Leads Us to believe it's hardware and/or BIOS Software
problem:

Case1: Two customers running Essbase on a IBM 720 server, writing data
to a RAID 5 storage device,
were getting Invalid Block Headers every time they ran their calc.
Essbase would notify that the calc
completed successfully when in fact the calc did not complete.

Resolution: Both customers upgraded their BIOS (traffic cop for video,
ram, motherboard and other
hardware) from version 11 to 12 and the problem went away.

Case2: An Essbase customer was getting the Invalid Block Header during a
calculation run on a Compaq
server writing to an external RAID storage device. His Essbase server
was also randomly receiving the
"NT Blue Screen of Death", which is normally a sign that the server
hardware is bad. The customer
replaced the current SIMMs (Ram Chips) with a new set of SIMMs and the
problem immediately went
away.

Case3: Customer had 3 servers from the same manufacturer. Two were
working fine and the third server
kept receiving the invalid block header problem. When they moved the
application that was having
problems onto one of the two "working" servers, the application did not
encounter the Invalid Block
Header error. Upon further investigation, it was discovered that the bad
server had a different model of
Controller Card than the two working ones. They swapped in a new
controller to the server experiencing
the Invalid Block Header error and the problem went away.

Case4: We had replicated 2 Invalid Block Header cases on an external
RAID 5 storage array in house
that went away after Storage Computer came in and upgraded our BIOS
level, reconfigured some SCSI
settings, and moved the drives into a workstation box which had new
controllers.

Other Pertinent Information:

   1.The XCP logs didn't help diagnose the problem
   2.One of the possible reasons why this is happening is that Essbase
writes its page boundaries
     differently than a relational database and the BIOS in possible
concert with the other hardware may
     be having problems moving the data correctly because of this. When
we need to write to the page
     file, we write it in different sizes as needed to avoid wasting
space. For example, we could write a
     page file that is 6K or greater or smaller to ensure that the
entire page is getting written out.
     Relational databases write to only 4K sizes. Every page size is 4K
and if there is partial data written
     out, the rest of the room could be wasted. i.e. The relational
source has to write a page file that is
     4.1K in size. The first 4K would be written to the 1st page file.
The remaining .1K would take up
     the entire 2nd page file and the rest of that file is wasted.
   3.Our Engineering Department has been working directly with Microsoft
Engineers on this issue for
     several months and they believe the problem is the result of a
hardware conflict between the server
     OS layer and any RAID storage device.

What You Need To Do

     Need to have the Customer get a very technical person who is good
with hardware. This person
     needs to be able to ascertain the version of the BIOS and
Controller Card. He/she will need to call
     up the vendor of the server to see if they are aware of any new
upgrades for the BIOS, driver levels
     for the Controller, and possibly to get a hold of additional RAM
chips to swap with the original
     RAM in the problematic server. He/she may also need to check with
the vendor of any noted
     technical problems related to their equipment.

Hope it help,
Marcelo Narvaja



Fri, 26 Apr 2002 03:00:00 GMT
 
 [ 4 post ] 

 Relevant Pages 

1. Essbase "Invalid Block Header" Problems

2. ReportSmith Black Blocked Headers and Footers

3. segment header block

4. Block header average size

5. Tablespace Header Blocks??

6. ** Block Headers **

7. undo header/block

8. mt_shm_malloc_segid bad block header

9. Assert Failed: Memory free block header corruption detected in mt_shm_free_pool 3

10. Assert Failed: Memory free block header corruption detected in

11. Memory block header corruption detected in mt_shm_free

12. Assert Failed: Memory block header corruption detected in


 
Powered by phpBB® Forum Software