Interesting hack :) But are all fails so catching? What about memory limit
exceeding
for example?
воскресенье, 13 января 2013 г., 10:49:02 UTC+4 пользователь George Snelling
написал:
>
> Sooo, this is a very simple solution that works for us for a single,
> light-duty production server running at joyent. It solves problem 1 just
> fine and problem 2 when the root cause is a bug in our app. It won't work
> for big deployments, or for any server resource outage. I doubt many folks
> who solve those problems at scale try to build identical systems that work
> on both windows and nix.
>
> We launch our production server using nodemon. We like nodemon because it
> is small, simple, and reliable. Using nodemon as a process runner, most of
> our server deployments are just git pulls.
>
> Then we have a fancier version of this function in our main:
>
> process.on('uncaughtException', function(err) {
> // make sure you don't have a startup problem
> require('fs').writeFileSync(__dirname + '/crash.js', err.stack||err)
> // page the ops guys
> process.exit(1)
> }
>
> On an uncaught exception, nodemon notices that the file crash.js changed
> and bounces the node process, restarting the server in a stable state. We
> originally held our nose writing this hack, planning to contract with some
> monit guru to write a better solution, but it has worked so well that we
> have become comfortable with it over time.
>
> Again, not a real solution for high-scale production apps, but for all the
> unwashed nodettes out there who value simplicity and want to sleep
> peacefully with a single-box server, this works pretty well.
>
> Cheers!
>
> -george
>
> On Saturday, January 12, 2013 3:17:08 PM UTC-8, nin jin wrote:
>>
>> Behaviour is similar, but different
>> * Service runs with system privelegies. This is dangerous.
>> * nssm is windows specific tool. Better to have one application stack on
>> all platforms.
>> * Аnyway someone must watch for files changes.
>>
>> 2013/1/13 José F. Romaniello <[email protected]>
>>
>>> I'd ask you the inverse, why would you reinvent the service function of
>>> windows or *nix daemons in node? It's there, it is easy to use, the OS
>>> comes with it. These things do a pretty good job starting the service when
>>> booting, and trying to restart when it exist according the exit status.
>>
>>
>>
>>
--
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
