We had some race conditions with ACK management which caused us to not
properly clean up some transactions in the send queue. My guess is that is
what you are seeing. If you bounced the server, they should go away.

If you want, you could also examine the progress tables to see if the
transactions have been acked from all of the targets or not. You could then
manually delete the old transactions, if you wanted. It gets a bit
complicated to explain how to do this over a news group and is somthing that
I'd rather not do.


"Heinz Weitkamp" <heinz.weitkamp@westfleisch.de> wrote in message
news:1127922414.07b51ac69a867a753d5b64ecea7fd181@t eranews...
>
> Hi all,
>
> IDS 7.31 UD4
> SCO Unix 3.2 5.0.6
>
> In the syscdr db are two tables "trg_receive_sbuf" and "trg_receive_stxn"
> which contains only on one server (we have 12 server in ER)
> many rows. The rows are very old (committime: 3 month). Can I delete the
> rows without any trouble with ER?
> If yes, must i stop and start ER?
> What are the meaning of the rows?
>
> Thanks in advance
>
> Heinz
> sending to informix-list