I have a coworker who is trying to run 6 bcp's at the same time into 6
different tables. My first thought was don't do things like that. But
I help my tongue, I'm getting better. :-)

The say that one bcp will work and the others will apparently block on
it. I did not see this myself, but I'm wondering why would that be?

Details

Sybase 12.5 on AMD 64bit. Solaris
Regular bcp into tables with indicies and no format file.

I can't think of any reason why there would be blocking. Unless some
system table is being locked or the transaction log hogged?

Any ideas anyone?

On Feb 25, 4:22*am, Wayne Happ <wayneh... (AT) earthlink (DOT) net> wrote:
If you're running logged bcp inserts then there will probably be some
pretty heavy IO, especially to the transaction log, but off-hand I
can't think of a system table that would be locked for the duration of
the bcp in.

I guess running sp_lock during a trial run would clarify things... :-)

Michael

On Feb 25, 6:28 am, "mpepp... (AT) peppler (DOT) org [Team Sybase]"
<michael.pepp... (AT) gmail (DOT) com> wrote:

It's certainly a strong possibility that the transaction log
is being filled up, depending of course on how much space is
allocated for it.

The database will come to a screeching halt if the transaction
log fills up. You can tell that with "sp_who", which will tell
if users are blocked waiting for transaction space.

What I have done in cases where I have large amounts of data
to bcp in is start another background process that does a
"dump transaction with truncate_only" every few minutes until
the bcp is complete. Then, when finished, do a full dump
and return to the normal transaction dump schedule.

/:-/