Corrupting Database

[News]

Hi all,

Here is my setup.

mdb backend on win03 server - roughly 80MB compacted - growing at 200kb/day
mde frontend on clients - roughly 28MB compacted
all clients run AccessXP
Half clients on NT4SP6 - Half clients on WinXPProSP1 - 1 Client on Win2KSP4
all clients run Jet 4.0 Sp8
max ca. 15 concurrent users

This setup has worked for us for the past 2 years, with little hitch

Over the past week, I have experienced 3 database corruption events.

Nothing new has happened, no new client plugged in, no software change or
hardware change.

Good news is that i can repair the backend mdb with the Repair and Compact
Tool.




THE 3 EVENTS

Event #1:
Tuesday around 10:30am.
Got everyone out of it
Repaired compacted, and back and running

Event #2:
Wednesday around 10:30am
Got everyone out of it
Repaired compacted, and back and running
Started peeing in my pants
Thought maybe had something to do with the time, started reading the
newsgroups
Applied the Jet SP8 to everyone, as I thought it would fix a concurrent
date sorting issue I had with one of my forms

Event #3
Thurday around 5pm
Got everyone out of it
Repaired compacted, and back and running
My pants are wet
I noticed though that for some reason, while some users could not connect
and got the unknown database format message, others were able to do their
thing for an additional 2-3 minutes.
After I repaired the mdb again, it appears that whatever those users wrote
in those noman'sland 2-3 mn, was actually written in the mdb
(Side Note, I found that my date sorting problem was due to one NT4SP6
machine having its regional settings date format messed up)


I have read most of the msft articles on corruption. My point is that the
fact that someone could still write in the "corrupted" mdb, for some few
minutes after, is addressed nowhere. I am not an MVP or anything but
wouldn't a corrupt mdb be corrupt for everyone at the same time? If not,
does this at least give a hint to anyone as to why my particular
corruption is happening?

I would love to solve this for next week, I don't have that many pairs of
pants.

Thanks all

d

 

[News]

Hi all,

Here is my setup.

mdb backend on win03 server - roughly 80MB compacted - growing at 200kb/day
mde frontend on clients - roughly 28MB compacted
all clients run AccessXP
Half clients on NT4SP6 - Half clients on WinXPProSP1 - 1 Client on Win2KSP4
all clients run Jet 4.0 Sp8
max ca. 15 concurrent users

This setup has worked for us for the past 2 years, with little hitch

Over the past week, I have experienced 3 database corruption events.

Nothing new has happened, no new client plugged in, no software change or
hardware change.

Good news is that i can repair the backend mdb with the Repair and Compact
Tool.




THE 3 EVENTS

Event #1:
Tuesday around 10:30am.
Got everyone out of it
Repaired compacted, and back and running

Event #2:
Wednesday around 10:30am
Got everyone out of it
Repaired compacted, and back and running
Started peeing in my pants
Thought maybe had something to do with the time, started reading the
newsgroups
Applied the Jet SP8 to everyone, as I thought it would fix a concurrent
date sorting issue I had with one of my forms

Event #3
Thurday around 5pm
Got everyone out of it
Repaired compacted, and back and running
My pants are wet
I noticed though that for some reason, while some users could not connect
and got the unknown database format message, others were able to do their
thing for an additional 2-3 minutes.
After I repaired the mdb again, it appears that whatever those users wrote
in those noman'sland 2-3 mn, was actually written in the mdb
(Side Note, I found that my date sorting problem was due to one NT4SP6
machine having its regional settings date format messed up)


I have read most of the msft articles on corruption. My point is that the
fact that someone could still write in the "corrupted" mdb, for some few
minutes after, is addressed nowhere. I am not an MVP or anything but
wouldn't a corrupt mdb be corrupt for everyone at the same time? If not,
does this at least give a hint to anyone as to why my particular
corruption is happening?

I would love to solve this for next week, I don't have that many pairs of
pants.

Thanks all

d

 

[John Vinson]

Over the past week, I have experienced 3 database corruption events.

Just a quick suggestion (well, two):

- Check out the suggestions at Tony Toews' excellent site

http://www.granite.ab.ca/access/corruptmdbs.htm

- Check out EVERYONE's NIC - network connectivity problems (noise,
dropped connections, etc.) are probably THE main cause of such
corruption. Is anyone connecting to the database using a wireless LAN?
If so... stop them. Wireless is usually not sufficiently stable to be
usable by the very demanding Access connection.

John W. Vinson[MVP]

 

[News]

Acutally, thanks John
I had checked Tony's site, and I very much agree that it is excellent.

I thought perhaps my particular case was indicative of a cookie-cutter
corruption case
anyway, I guess I will have no choice but to check the NICs

My problem is that I tried to send back and forth a huge zipped file, and
the NICs seem to be fine

Somehow, I guess, they fail or are weak once in a while. It just seems
hectic to replace ca 20 pc NICs.

Anyway, FYI, the network is solid state, i.e. no wireless nothing, so I
guess this has to be disregarded as a possible cause.

Again thank you, and let me know if you think of something else.

d

 

[John Vinson]

Again thank you, and let me know if you think of something else.

Corrupted tables - in particular corrupted memo fields - can cause
wierd problems. Try creating a new database for the backend and
importing all the tables.

John W. Vinson[MVP]