> This is an error I am getting with 6.3.2
> under freebsd 2.2.8
> Mon Sep 6 16:44:48 BST 1999: NOTICE: SIAssignBackendId: discarding tag 2147483020
> Mon Sep 6 16:44:48 BST 1999: FATAL 1: Backend cache invalidation initialization failed
>
>
> Our database is about 38megabyes. We do have about 20 - 30 simultaneous connections. Most accessing the same
> tables with transactions.
>
> We see the above error quite often, and once it happens we have to restart postgress and our application.
> We are accessing the db via jdbc.
>
> We do not use any large objects (since they caused constant backend crashes for us)
>
> 1) Can anyone explain what this error is, and if there is something we can do to work around it.
It is caused by the corrupted shared cache. No workaround exists for
6.3.2 as far as I know.
> 2) is 6.5 stable enough for 23.5x7 production applications.
6.5 is much stable than pre 6.5 including 6.3.2. Even 128 simultaneous
connections are fine if properly configured.
> 3) are large objects stable in 6.5 (where I can store and access 20,000 of them regularily)
Yes, except that 20,000 large objects would be slow (20,000 large
objects will create 40,000 files right under the database directory
that would make directory lookup slow). If you are serious about using
many large objects, probably I could supply patches to enhance the
performance as a workaround.
> 4) if all my sql works in 6.3.2 will it need any changes to run under 6.5
It depends on your sql works. But basically you should be required
very few changes, I believe.
> I am loosing the -- flush this free crap and move to oracle -- war here, so please help.
I'm getting wining over commercial DBMSs in many projects since 6.5
released. Thanks for the row-level locking and MVCC stuffs..
---
Tatsuo Ishii