Re: Server job locked up because a lock wait timed out. I vaguely remember that there's something I can do, . . . .

Hi, James:

Was that program doing a READE or READPE when it encountered the record lock situation? If so, you can put an indicator in the "LO" resulting indicator position, and if a record lock is the problem, RPG will turn on that indicator rather than throwing an exception that ends up in your error handling subroutine. Or, you can issue a READE(N) or READPE(N) to do the read with no locking (if your application can tolerate that.) If you are reading the record prior to updating it, you would need to read the record again with a record lock.

HTH,

Mark

> On 9/9/2010 7:32 PM, James H. H. Lampert wrote:

I wrote:

I've just had an ILE RPG exit program lock up a server job on a customer
site, because it was kept waiting too long on a lock.

I can't recall ever seeing this happen before, but we all know that for
every case when something like this gets reported, there's probably
several that don't.

I have this vague memory that there's something I can do about this
situation, to keep it from locking up a user's server job, but I can't
remember what.

It's a record lock, of course. And no, I no longer have a joblog to excerpt.

Isn't there either a way to have a program wait indefinitely for a
record lock, or a way to have it catch the exception when a lock wait
times out?

--
JHHL