Re: Debian MySQL Perl DBI - connection terminates unexpectedly after 100 secs.


"Jens Thoms Toerring" <jt@xxxxxxxxxxx> wrote in message
news:727ga0Fo7qv3U1@xxxxxxxxxxxxxxxxxxxx
John <john1949@xxxxxxxxx> wrote:

"J. Gleixner" <glex_no-spam@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:49be71fd$0$89873$815e3792@xxxxxxxxxxxxxxxxx
John wrote:
I'm updating a MySQL table with about 2 million records. After about
100
seconds the program ends with no error message. If I use LIMIT
800000,50000 the part where it stopped is OK so the UPDATE part is OK.

Is there a time out parameter? Or am I running out of memory. Can
the
memory be increased?

Check line 24.

Line 24 says:-

our $old_fh=select(STDOUT); $|=1; select($old_fh); # Make standard output
socket hot

Can't see how that can affect it.

I rather guess it was a gentle hint that your problem description
is too vague to allow to deduce what's going wrong. But your sub-
ject line may contain a hint: it looks as if the update is done
from within a CGI script or something similar, running on a web
server. And normally there's a CPU time limit set for such scripts
to catch e.g. script that run in an endless loop. That's what I
would check for first (note: if it's not set then an OS default
limit may be used!). If you are using Apache it might be a RLimitCPU
directive. But there could also be a limit on the memory consumption
of such scripts (RLimitMEM with Apache2)... Perhaps a look at the log
files of the web server will give you some more hints.

Regards, Jens
--
\ Jens Thoms Toerring ___ jt@xxxxxxxxxxx
\__________________________ http://toerring.de

Many thanks. That was very helpful. I have maxed both RLimitCPU and
RLimitMEM in Apache2 and that has had led to better performance. There are
severe problems in using MySQL when you start hitting more than 2 million
rows irrespective how well you tweak your settings.

Regards
John



.

Relevant Pages