Hi All,

I'm running Pervasive.SQL V8.5 on Redhat 9. My application runs for a
while and then suddenly the application has stop. And are getting
status 116 I've check the /var/log/messages I'm seeing this error
"mkded: LsnThrd: beginthread(SelectThread) error=12".

Can anyone help on this one?


thanks

jhun

Sounds like it could not start a thread.

The top reasons I know of would be:
1) out of memory to start a thread. Double check the memory settings.
2) trying to start too many threads. Double check the thread
settings.

Since it is after a few days I would suspect the first more than the
second. May just want to back off the Max Micro Kernel Memory Usage
setting to 40 or even 0.

You should be able to see how much memory is being consumed in the top
monitor utility. A ps -ef will list the threads (actually pstree may
be easier). Check those periodically and note if one or the other
rises over time.

Leonard

On 18 Feb 2004 20:10:32 -0800, jhun.bacala (AT) nccc (DOT) com.ph (jhun bacala)
wrote:

Status 116 is usually a Workgroup engine message and, though you
weren't specific about which Pervasive.SQL V8.5 you are running, one
would assume it is the server engine if it is running on Redhat 9,
I've never seen it from a server engine. We all know why one has to be
careful about assuming things without enough specfic accurage
information though :-) You also didn't say what you application was
running on, but I'm guessing it's on Windows so if I were you I would
check for a workgroup engine running on that PC and turn off the local
access so the client, and also the application, is forced to go
through the server engine via the requester rather than use the
workgroup engine.

Please excuse me if your application is a Linux app. If the error=12
is the same as a Status 12, then it could be a pathing issue of some
sort since it can't find the file.

You can download and run PSA to detect if a workgroup engine is local
the the PC the application runs on.

Regards,
JR

Hi John,

I'm very sorry for lack of information actually, I'm running
pervasive.sql v 8 server on redhat machine. My application is on
windows and we're using Magic software.

I've tried lowering the Max Microkernel Memory Usage from 90 to 50.
Also their are few users now accessing that server. And now I'm not
seeing the above error. One thing I've notice in version 8 is that I'm
only seeing one mkded process, compared with 2000i that spawns a lot
of processes. Are all the processes being share with only one mkded
process?

Btw, I'm taking the advice of Leonard of lower the Max Microkernel
Memory Usage to 40 or 0, what's your take on this one?

Thanks,

Jhun Bacala
jrothgeb (AT) pervasive (DOT) com (JohnR) wrote in message news:<bf1688a9.0402191011.21a1b68f (AT) posting (DOT) google.com>...

RE: Thread usage.
The Pervasive engine uses about the same threads for V8 as it did for
2000i. Highly dependent on settings.
The one thread v. many probably has more to do with which version of
Linux you are running. Starting with about RH 7 they started showing
all of them as just one thread under ps. Doing a pstree it will show
you the counts.

Leonard

On 19 Feb 2004 19:10:40 -0800, jhun.bacala (AT) nccc (DOT) com.ph (jhun bacala)
wrote: