[Glass] Loading application time via Gofer or directly via topaz

[Dale Henrichs]

Marten,

Depending upon how you run your session there should/could be log files
that provide more detail about what has happened when the putty session
died ... at a minimum if you do a `output push topaz.log` you will have a
file on the server that you can look at after the putty session died to
help determine what might have happened ...  I would be suspicious that
perhaps the linux login is being killed by a linux OOM killer that is
activated when you are running critically low on system memory ... I don't
have the details at my finger tips, but I think that the oom killer logs
the pids of the processes that killed and from the topaz.log you should be
able to determine the pid of the topaz session and go from there ...

Dale

On Wed, Mar 2, 2022 at 9:18 AM Marten Feldtmann via Glass <
glass at lists.gemtalksystems.com> wrote:

> I do no get any error. I connect to the server via putty, start a topaz
> script, loading the Monticello package ... then it runs for about 15-20
> minutes and then the process is gone, but the putty terminal still shows
> the process - but I assume, that putty has actually lost the connection.
> The shell process is not available any more.
>
> Perhaps I have to look at ssh configuration values, that ssh is
> terminating the session ...
>
> Marten
>
> Yanni Chiu <yanni.code at gmail.com> hat am 02.03.2022 17:11 geschrieben:
>
>
> What is the error/problem you are getting? I recall there being a problem
> in that the method source code can get mixed up with the syntax of topaz.
> The details are fuzzy to me now, but I think the issue was method source
> that looks like a topaz comment line.
>
> —
> Yanni Chiu
>
> On Mar 2, 2022, at 4:10 AM, Marten Feldtmann via Glass <
> glass at lists.gemtalksystems.com> wrote:
>
> Hey,
>
> so I decided to go away from Monticello packages for these very large
> source code files and switch to pure topaz code - the process does not seem
> to be reliable when the stuff is getting too large. I had to reinstall
> Gemstone database and got again the same problems - topaz code is loading
> much faster and the loading has ever worked ...
>
> Marten
>
> Marten Feldtmann via Glass <glass at lists.gemtalksystems.com> hat am
> 27.02.2022 18:25 geschrieben:
>
>
> Hey,
>
> I am currently working with Gemstone/S 3.6.3 and today I had lots of
> problems, loading large software packages via Gopher - the topaz process
> (using Gopher) got killed (as it seems) by the operating system when
> loading the software, no message and the putty session got killed also.
>
> I am heavily model based working, so the complete domain model, the API
> model and APIs itselfs, the OpenAPI specification, the documentation are
> generated in Smalltalk source code. My - by far - largest application
> produces 18 MByte of source code (topaz-pure code).
>
> The generator generates topaz source code (18 MB) which is normally loaded
> once and then (from within Jade) I produce a Monticello package (3.6 MB)
> and for further work I use these Monticello packages.
>
> Today I measured the loading time: 6 minutes (topaz) against 24 minutes
> (monticello) on the same machine - I assume, that a git based solution will
> not be much better, perhaps even worse ?
>
> Just for your information purposes ....
> _______________________________________________ Glass mailing list
> Glass at lists.gemtalksystems.com
> https://lists.gemtalksystems.com/mailman/listinfo/glass
>
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> https://lists.gemtalksystems.com/mailman/listinfo/glass
>
> _______________________________________________
> Glass mailing list
> Glass at lists.gemtalksystems.com
> https://lists.gemtalksystems.com/mailman/listinfo/glass
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/archives/glass/attachments/20220302/8487a643/attachment.htm>