We're running multiple jobs simultaneously, which internally do kindly the same.
(Which means, they always perform queries on the same tables, in the same order.)

Example:

set lock mode to wait
set wait time to 180 (which is quite long, imho)
select table a for update...
update something in table a...


Now, if there are (let's say) 20 jobs doing the same thing (with different data content, of course) simultaneously, Informix frequently returns ISAM-Error -143 (which is: "deadlock detected").

From my point of view, deadlocks are not possible in this kind of processing.

As i ascertained, Informix queues requests internally, processing them one by one.
Is there a limit of concurrent requests that Informix can handle?
If so, is there a environment variable one can set?

Looks like Informix gets mixed up handling a certain amount of concurrent requests.

Thanks for comments and proposals...

which mode does your table set ? (page,row)
My site was encountered this problem a month ago and have bee solved,
Can you set value of OPTCOMPIND parameter to 0 and run your command again (table lock mode = row)

Our tables are set to row-lock.

I will try this option as soon as i reach a db-admin.
I'll inform you on the results.

Thanks!