third_party_content:community:commentary:forums_nonworks:b0448e1f8e5e01a1852569d700725b40

Sign up on the Revelation Software website to have access to the most current content, and to be able to ask questions and get answers from the Revelation community

At 17 JAN 2001 03:49:04PM Ted Archibald @ parsec systems wrote:

To all you experts with loads of experience;

here is one to sink your teeth into.

One of my clients has a sales history file SALES

which has 7.4 Mrec, LK=205 MB OV=350 MB.

Many times AREV report pgms that use the SALES file

will crash with a GFE but when I use DUMP

to view (nnnnn CTL+G) the group there usually (98%)

is no GFE indicator at the bottom of the

DUMP screen, thus a Phantom GFE. Also, the

LH_VERIFY does not find them. 2% are solid GFEs.

There have seen several instances where two

programs running simultaneously have encountered

the same GFE and at least one instance where

two subsequent programs encountered the same GFE.

This indicates to me that the GFE exists at least

in the Server memory.

In all cases the DUMP and other reports that

repeat the access to the same records in the SALES

file report no GFE.

The occurance of Phantom GFE's is about 1/day or

about 1/10 reports that access the SALES file.

Needless to say that the users and their management

are beginning to question the stability of the system.

No other files are effected.

In the past I have copied the file record/record

to another file and did a re-name

and this did not solve the problem.

Here is my analysis so far:

1-the network seems not to be the problem

   (nics, cables, routers, hubs, etc)
   I have done file copy compare for days on the
   suspect workstations and find nothing.

2-Hardware Memory

   Could the problem be in the memory of the Server
   (Novell 3). Suppose that there was a memory problem
   or a disk controller problem that corrupted a
   LH block on the read (occasionally)
   Data gets staged to memory where it
   would remain for short time and then purged.
   This would give programs running in the same time
   frame the same data and same GFE.  If the data was
   read only then the error would be flushed by Novell.
   If the data in the Novell block and not the LH group
   was updated and written back then the GFE would be
   now permanent.   If the data was in the damaged LH
   group then AREV would signel with a GFE.
   This would account for the 5% or so
   permanent GFEs to the file
   I suspect that hardware memory is the problem.

Config=Novell 3.12, Arev 3.12 , server Pentium I, 400MHz 300 MB

Arev NLM 1.5, 2xAdaptec controllers,2xSeagate 8 GB (mirrored)

Any suggestions as to what I might do to solve this problem.

or what other tests might I perform to get closer to the solution.

Cheers

Ted @ parsec@direct.ca


At 18 JAN 2001 02:47AM Stefano Cavaglieri wrote:

Phantom GFEs, in a Novell 3.x-4.x environment, is quite a common (and a serious) issue if you work with large files and the "turbo FAT" is enabled on the server side (it is by default). If this is your case, try getting a fix from Novell. If I remember well, the name of the patch for NetWare 3.12 was FATFIX.

Hope it helps,

Stefano


At 18 JAN 2001 07:19AM [url=http://www.sprezzatura.com" onMouseOver=window.status=Click here to visit our web site?';return(true)]The Sprezzatura Group[/url] wrote:

Stefano's seems like the best solution (Hi Stefano - happy new year! ), but if this is not the problem check the workstations to ensure that write behind cacheing is disabled on the workstation. (Right click on My Computer, Properties, Performance, Toubleshooting, File System).

The Sprezzatura Group

World Leaders in all things RevSoft

View this thread on the forum...

  • third_party_content/community/commentary/forums_nonworks/b0448e1f8e5e01a1852569d700725b40.txt
  • Last modified: 2023/12/28 07:40
  • by 127.0.0.1