Smalltalk › Frameworks & Tools › Seaside › Seaside General

Seaside image health check

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

6 messages Options

Paul DeBruicker

Seaside image health check

I have a Pharo 1.3-13315/Seaside 3.0.6.3 headless image on 64bit linux
that sometimes stops responding to HTTP requests and acts like its in an
infinite loop. If I connect to that image with the RFB after a couple
minutes I can enter my password and it drops out of the infinite loop
and resumes answering HTTP requests as if nothing was wrong. The hang
happens irregularly and I haven't yet been able to get access to a
debugger when it is hanging. CPUWatcher doesn't catch it either.

I'm thinking to use the Scheduler package + ZnClient to send a request
from that image to itself requesting a page, and if it times out, to
write the ProcessBrowser + CPU usage data to a log file.

Is there an established/better/other way to attempt to diagnose these
intermittent hangs?

Thanks

Paul
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Philippe Marschall

Re: Seaside image health check

2011/12/22 Paul DeBruicker <[hidden email]>:

> I have a Pharo 1.3-13315/Seaside 3.0.6.3 headless image on 64bit linux that
> sometimes stops responding to HTTP requests and acts like its in an infinite
> loop. If I connect to that image with the RFB after a couple minutes I can
> enter my password and it drops out of the infinite loop and resumes
> answering HTTP requests as if nothing was wrong. The hang happens
> irregularly and I haven't yet been able to get access to a debugger when it
> is hanging. CPUWatcher doesn't catch it either.
>
> I'm thinking to use the Scheduler package + ZnClient to send a request from
> that image to itself requesting a page, and if it times out, to write the
> ProcessBrowser + CPU usage data to a log file.
>
> Is there an established/better/other way to attempt to diagnose these
> intermittent hangs?

If the RFB fixes it it's may be the opposite of an infinite loop.
Nothing happening because of a Socket/Semaphore/ProcessScheduler
issue. The next time it happens can you check the CPU load with top to
make sure?

Cheers
Philippe
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Paul DeBruicker

Re: Seaside image health check

On 11-12-22 07:39 AM, Philippe Marschall wrote:

>
> If the RFB fixes it it's may be the opposite of an infinite loop.
> Nothing happening because of a Socket/Semaphore/ProcessScheduler
> issue. The next time it happens can you check the CPU load with top to
> make sure?
>
> Cheers
> Philippe
> _______________________________________________
>

Thanks Phillipe. The times Ive noticed the hang I've checked CPU use
with htop while waiting for the RFB login window to appear and the CPU
has been maxed out by the image. I'm sure its something I've written
that's bad, I'm just not sure what/where.

FWIW, the hang is happening right now. It started again about 15mins
ago. The Squeak process is at 94%. The load average is 1.73 1.56
1.14. When I start the image I use the -mmax=256MB, and htop shows its
only using 100mb and its steady. I've been logging the number of
instances of my custom session each time a new one is created and it
shows that there are only 6 right now.
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Levente Uzonyi-2

Re: Seaside image health check

In reply to this post by Paul DeBruicker

On Thu, 22 Dec 2011, Paul DeBruicker wrote:

> I have a Pharo 1.3-13315/Seaside 3.0.6.3 headless image on 64bit linux that
> sometimes stops responding to HTTP requests and acts like its in an infinite
> loop. If I connect to that image with the RFB after a couple minutes I can
> enter my password and it drops out of the infinite loop and resumes answering
> HTTP requests as if nothing was wrong. The hang happens irregularly and I
> haven't yet been able to get access to a debugger when it is hanging.
> CPUWatcher doesn't catch it either.
>
> I'm thinking to use the Scheduler package + ZnClient to send a request from
> that image to itself requesting a page, and if it times out, to write the
> ProcessBrowser + CPU usage data to a log file.
>
> Is there an established/better/other way to attempt to diagnose these
> intermittent hangs?

If you're using CogVM, then you can force printing the content of all
smalltalk stacks to stdout by sending a USR1 signal to the process with
the following command:

kill -USR1 <pid>

The currently active process will be printed first.

Levente

>
> Thanks
>
> Paul
> _______________________________________________
> seaside mailing list
> [hidden email]
> http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside
>
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

Paul DeBruicker

Re: Seaside image health check

On 11-12-22 08:46 AM, Levente Uzonyi wrote:

>>
>> Is there an established/better/other way to attempt to diagnose these
>> intermittent hangs?
>
> If you're using CogVM, then you can force printing the content of all
> smalltalk stacks to stdout by sending a USR1 signal to the process
> with the following command:
>
> kill -USR1 <pid>
>
> The currently active process will be printed first.
>
>
> Levente

Thanks! I'll try that next time it happens.
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

sebastianconcept@gmail.co

Re: Seaside image health check

In reply to this post by Levente Uzonyi-2

BTW, I've tried a couple of times CogVM but always failed to get it stable.

Who can make use of the dump logs?

sebastian

On Dec 22, 2011, at 2:46 PM, Levente Uzonyi wrote:

On Thu, 22 Dec 2011, Paul DeBruicker wrote:

I have a Pharo 1.3-13315/Seaside 3.0.6.3 headless image on 64bit linux that sometimes stops responding to HTTP requests and acts like its in an infinite loop. If I connect to that image with the RFB after a couple minutes I can enter my password and it drops out of the infinite loop and resumes answering HTTP requests as if nothing was wrong. The hang happens irregularly and I haven't yet been able to get access to a debugger when it is hanging. CPUWatcher doesn't catch it either.

I'm thinking to use the Scheduler package + ZnClient to send a request from that image to itself requesting a page, and if it times out, to write the ProcessBrowser + CPU usage data to a log file.

Is there an established/better/other way to attempt to diagnose these intermittent hangs?

If you're using CogVM, then you can force printing the content of all smalltalk stacks to stdout by sending a USR1 signal to the process with the following command:

kill -USR1 <pid>

The currently active process will be printed first.

Levente

Thanks

Paul
_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside

_______________________________________________
seaside mailing list
[hidden email]
http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside