O
Ondine
Hi
I have a client with a replicated back-end database (Ac 2000/win xp) who
sporadically suffers from data corruption. The last incident was typical:
the client was working in a record (main table and linked sub-records) and
received a 3167 'Record is deleted error', all the fields showed 'Error' and
he was unable to exit the form and had to reboot. After this they compacted
and repaired the back-end data file on the server but afterwards they found
themselves in an old backup of the data from a bak file and the file they had
been using had been renamed 'db1_damaged.mdb' (is this something which only
happens on replicated databases?).
The db1_damaged file was no longer a replicated database, but upon
examination I could find no evidence of any corruption in that particular
table nor in any others, and the file could be successfully
compacted/repaired. I created a new back-end data file and transferred the
data and also created a brand new table for the table which had caused the
problem and appended the data into the new table.
This particular table also sometimes seems to have an indexing problem
whereby the client will go to open a record and instead another record will
open. The record which opens in error is always the same record and I
believe it is the first record in the table. However, this happens only from
time to time. Unfortunately it is still happening even though I created the
new table (without replication fields) before converting the new back end
into a replicated db. There must be some minor corruption in the data but I
can't find it. There are no memo fields, no hyperlinks, and each column can
be sorted and queried.
A major corruption 'meltdown' only happens once or twice a year, but that is
still too often as it frequently leads to me having to spend a huge amount of
time manually synchronizing data changes and also causes serious downtime for
the client. The problem has occurred when the client was working on his
local replicated data file and so I doubt it is a network issue. There are
usually no more than 2 or max 3 users on this system.
I am seriously considering typing out all 620 records if I thought it might
solve the problem! The sub-table has more than 5000 unfortunately.
Does anyone have any ideas as to how I might be able to find the source of
this problem and steps I could take to solve it?
Many thanks.
Ondine.
I have a client with a replicated back-end database (Ac 2000/win xp) who
sporadically suffers from data corruption. The last incident was typical:
the client was working in a record (main table and linked sub-records) and
received a 3167 'Record is deleted error', all the fields showed 'Error' and
he was unable to exit the form and had to reboot. After this they compacted
and repaired the back-end data file on the server but afterwards they found
themselves in an old backup of the data from a bak file and the file they had
been using had been renamed 'db1_damaged.mdb' (is this something which only
happens on replicated databases?).
The db1_damaged file was no longer a replicated database, but upon
examination I could find no evidence of any corruption in that particular
table nor in any others, and the file could be successfully
compacted/repaired. I created a new back-end data file and transferred the
data and also created a brand new table for the table which had caused the
problem and appended the data into the new table.
This particular table also sometimes seems to have an indexing problem
whereby the client will go to open a record and instead another record will
open. The record which opens in error is always the same record and I
believe it is the first record in the table. However, this happens only from
time to time. Unfortunately it is still happening even though I created the
new table (without replication fields) before converting the new back end
into a replicated db. There must be some minor corruption in the data but I
can't find it. There are no memo fields, no hyperlinks, and each column can
be sorted and queried.
A major corruption 'meltdown' only happens once or twice a year, but that is
still too often as it frequently leads to me having to spend a huge amount of
time manually synchronizing data changes and also causes serious downtime for
the client. The problem has occurred when the client was working on his
local replicated data file and so I doubt it is a network issue. There are
usually no more than 2 or max 3 users on this system.
I am seriously considering typing out all 620 records if I thought it might
solve the problem! The sub-table has more than 5000 unfortunately.
Does anyone have any ideas as to how I might be able to find the source of
this problem and steps I could take to solve it?
Many thanks.
Ondine.