On 5 March 2012 19:49, chris.percol <[email protected]> wrote:
> I look after a private cloud of around 30 windows and linux virtual
> servers.
>
> We are using node to do some pretty cool stuff at the moment on these
> servers e.g. run sql commands, file backups and other stuff. However,
> I'm individually pulling code changes down to servers using git which
> is far from ideal.
>
> What would folk recommend for better deployment of code updates? In
> particular something that is clever enough to know if node is
> processing a job (we use kue at moment) that the autodeploy doesn't
> kill the running jobs.
>
> Many thanks,
>
> Chris

I had a thought that one could probably re-require modules without
restarting the node process by sending a signal or something like
that. Also, I guess setting file watchers would be a solution also and
you could even implement signature checking or something like that I
suppose. I would appreciate to hear from anyone who knows V8 better
whether it would be posible to re-require a module and if yes are
there any limitation/downsides  in doing such kind of thing?

Another thing would be,if you are using node for some sort of work
queues, you could probably make this with watcher-supervisor kind of
process that will signal it's children to exit once the last job is
completed and respawn those with the updated code. This way you are
not respawning for each job, i.e. workers sit idle if there is no
work, but exit only when they are told to do so in order to run of the
new code base.

Cheers,
-- 
Ilya

-- 
Job Board: http://jobs.nodejs.org/
Posting guidelines: 
https://github.com/joyent/node/wiki/Mailing-List-Posting-Guidelines
You received this message because you are subscribed to the Google
Groups "nodejs" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/nodejs?hl=en?hl=en

Reply via email to