Search found 1 match

by hbus
Wed 31 Aug 2011 08:46
Forum: dotConnect for PostgreSQL
Topic: PostgreSQL arrays get corrupted after a while
Replies: 2
Views: 1253

PostgreSQL arrays get corrupted after a while

We have recently changed to use protocol=3 and are experiencing some fatal issues.

We have an aggregation process running that performs a lot of queries (and some updates) for hours. After a while - often between one and two minutes - we are having issues with columns that are returned from PostgreSQL as arrays. The driver never seems to recover and will afterwards always produce incorrect array columns.

We are using DataTables and are relying on the driver's ability to convert the PostgreSQL array to a string in the format of "{abc,def}". That has been working great for a long time using protocol=2.

The string we get when it gets corrupted is such as: "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\rabc\0\0\0\rabc"

However, restarting the application and running the same query again produces the correct result.

I have been trying hard to find a way to reproduce the issue without our huge dataset but have so far been unable to do that.

We are experiencing the issue using PostgreSQL 9.0.4 on both Windows and Linux and with both the 4.95.190 and 5.30.202 driver versions.

Any ideas on how to gather additional details from the driver on this issue?