# Deleted in memo fields

[JJ]

I have an Access 2000 DB which has been in continuous use for six
years. It has a front/back configuration, I use a Linux box as a back
end file server, the DB has at most 10 simultaneous users.

We recently upgraded parts of our network to 1 GB. I also upgraded to
a new Linux box with Fedora 7 installed (from Red Hat 7). The HD is a
Seagate SATA drive.

Prior to these changes, the system was very stable. I would rarely
have to repair any of the back end files. In fact I went over a year
without doing any repair for corruption, I would occasionally just do
one to shrink the file size.

The number of records in the largest back end file is ~ 30,000. Most
of the DB have between 5,000 and 10,000 records. There are 26 DB.

My problem is that on an almost daily basis several records are
corrupted. It is almost always in a memo field, and almost always in
one of three or four memo fields. There are many memo fields that
never have a problem, and a number of the data bases continue to
function perfectly. When a memo field in the affected DB shows the
dreaded "# Deleted", the text and numerical fields are not affected.

Does anyone have an idea of why this is happening? We've replaced a
number of network components (switches, NIC, cabling) to see if it
could be an intermittent connection problem. The problems continue.

Is it possible that the DB are corrupted beyond the capabilities of
the Compact and Repair function? Do I need to copy all the data out to
another format, and repopulate the DB? Does the higher network speed
contribute to the problem? Should I go back to RedHat 7? I'd
appreciate any ideas on what is happening and how I can fix it.

Thanks

Jack Jacobson

 

[Armen Stein]

I have an Access 2000 DB which has been in continuous use for six
years. It has a front/back configuration, I use a Linux box as a back
end file server, the DB has at most 10 simultaneous users.

We recently upgraded parts of our network to 1 GB. I also upgraded to
a new Linux box with Fedora 7 installed (from Red Hat 7). The HD is a
Seagate SATA drive.

Prior to these changes, the system was very stable. I would rarely
have to repair any of the back end files. In fact I went over a year
without doing any repair for corruption, I would occasionally just do
one to shrink the file size.

The number of records in the largest back end file is ~ 30,000. Most
of the DB have between 5,000 and 10,000 records. There are 26 DB.

My problem is that on an almost daily basis several records are
corrupted. It is almost always in a memo field, and almost always in
one of three or four memo fields. There are many memo fields that
never have a problem, and a number of the data bases continue to
function perfectly. When a memo field in the affected DB shows the
dreaded "# Deleted", the text and numerical fields are not affected.

Does anyone have an idea of why this is happening? We've replaced a
number of network components (switches, NIC, cabling) to see if it
could be an intermittent connection problem. The problems continue.

Is it possible that the DB are corrupted beyond the capabilities of
the Compact and Repair function? Do I need to copy all the data out to
another format, and repopulate the DB? Does the higher network speed
contribute to the problem? Should I go back to RedHat 7? I'd
appreciate any ideas on what is happening and how I can fix it.

Thanks

Jack Jacobson

We once had very bad corruption problems with an MDB that (without us
knowing) had been placed on a Unix server. Some operating systems
don't handle the OS record locking scheme that Access uses. Other
than unrepairable corruption, another symptom was that the LDB file
that appeared in the folder when the first user logged in, actually
DISAPPEARED when any subsequent users logged in.

Perhaps your new version of Linux can't do it. Try moving the
back-end mdb to a Windows box and see if the corruption stops.

Armen Stein
Microsoft Access MVP
www.JStreetTech.com