[Glass] Cannot stop Seaside gems anymore

[Dale Henrichs via Glass]

On 12/21/15 2:55 PM, Mariano Martinez Peck wrote:
>
>
> On Mon, Dec 21, 2015 at 7:37 PM, Dale Henrichs 
> <dale.henrichs at gemtalksystems.com 
> <mailto:dale.henrichs at gemtalksystems.com>> wrote:
>
>     It looks like the log is before you tried the kill? I don't see
>     the message about the SIGTERM being handled ...
>
>     If I wasn't clear, I wanted the stacks and log file from after you
>     had done a kill ...
>
>
> I can't. Just more info: the kill actually happens as the process of 
> the gem is finally killed (that's why I cannot attach to gdb.. "No 
> such process").
> I THOUGHT it was not killing the the gem because I could still 
> navigate the app. However, it would look like if kill error would 
> happen AFTER the actual kill, and so the stop seaside gems exits and 
> let's the rest of the gems alive, and this is why I saw the app still 
> running. Sounds reasonable?
>
> Thoughts?
>
okay .... so the Seaside gems that are getting sent the SIGTERM, but not 
all Seaside gems are getting sent SIGTERM?

It sounds reasonable that an error occuring somewhere in between kills 
could terminate things ... I'm not sure which seaside gem kill script 
you are using, but presumably it gets launched by #performOnServer: and 
in 3.2.9 we provide quite a bit of information about the failure, 
including raising an ExternalError that has the command, error message, 
errno and status from the command ... which should be enough to figure 
things out ....

Raising an error may be new behavior for the shutdown code ... I create 
GemServer>>performOnServer:status: so that I could add custom behavior 
on failure and a similar thing might be needed for your shutdown code ...

Dale


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.gemtalksystems.com/mailman/private/glass/attachments/20151221/4da86eb4/attachment.html>