Well, I think you could check the logs. Server couldn't just die without saying anything at all.
On Fri, May 23, 2014 at 12:25 PM, Sai <[email protected]> wrote: > Thank You Fedor for the quick reply. I'm trying to implement bud. > > However, I'm already using 0.11.14 pre version. However it is not upto the > mark on a dedicated server with 16 cluster workers (i.e. on an 8 core > machine). The workers kind of die after every 12-18 hours using it. Every > time Ill have to kill the process and start it again after every 12-18 hrs, > when number of workers come down. > > I'd like to know, how can I inspect what is causing this cluster process > to die? > > Thank You > Sai > > > On Thursday, May 22, 2014 9:31:19 PM UTC+5:30, Fedor Indutny wrote: > >> Hello! >> >> You could try setting `bud` (https://github.com/indutny/bud) in front of >> your servers, or >> using latest development version of node.js (0.11.x). Both have much >> better performance >> than node v0.10 >> >> Cheers, >> Fedor. >> >> >> On Thu, May 22, 2014 at 7:37 PM, Sai <[email protected]> wrote: >> >>> >>> Hi, >>> >>> I have implemented a proxy server based on spdyproxy(https://github.com/ >>> igrigorik/node-spdyproxy). I have added cluster on the binary file i.e. >>> bin/spdyproxy (http://pastebin.com/BvtpK6kc). >>> >>> I have tested this on various machines with node.js v0.11.14-pre >>> (downloaded this for round robin on clustering, as earlier it was not using >>> up all the workers) >>> >>> 1. Digitalocean droplets (dual core) >>> 2. a virtual server (16 cores cpu) >>> 3. a dedicated server (8 cores cpu) >>> >>> I'have opened this proxy now only for around 100-200 users anytime users >>> and I want to scale it till upto 1000 or even more users. This I have >>> bundled into chrome as an extension where users enable this and anonymise >>> themselves for privacy reasons. I have talked to the developer of spdyproxy >>> , he said more cpu and more RAM would not help as there are not compute >>> tasks, but I have taken number of more cores just for cluster module to >>> create the worker process. >>> >>> I have done tcp/ip tuning as well increased here is the sysctl.conf >>> additional tcp settings(http://pastebin.com/x9FnqY4u), files are >>> unlimited, following this suggestion(http://blog. >>> caustik.com/2012/08/19/node-js-w1m-concurrent-connections/) The core >>> problem on the dedicated server (which is production), the workers keep >>> dying after every 12-18 hours periods. >>> >>> I have been looking for what the problem is...! there is issue on tls >>> on this blog (https://blog.indutny.com/0.benchmarking-tls), is this >>> what causing the problem, or the server.js errors in the code >>> node-spdyproxy or is it that or still node.js not ready for this. >>> >>> Also what caused the threads(cluster process) die slowly one after the >>> other? >>> >>> How can I get rid of --debug-port for cluster in this build? >>> >>> Thank You >>> Sai >>> >>> -- >>> Job board: http://jobs.nodejs.org/ >>> New group rules: https://gist.github.com/othiym23/9886289#file- >>> moderation-policy-md >>> Old group rules: 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 unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To post to this group, send email to [email protected]. >>> >>> To view this discussion on the web visit https://groups.google.com/d/ >>> msgid/nodejs/a72d6692-5063-457a-bf2f-fc5595c3fd98%40googlegroups.com<https://groups.google.com/d/msgid/nodejs/a72d6692-5063-457a-bf2f-fc5595c3fd98%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- Job board: http://jobs.nodejs.org/ New group rules: https://gist.github.com/othiym23/9886289#file-moderation-policy-md Old group rules: 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 unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/nodejs/CAEv2VfLR-tkg6uoA2FAPvnK-B1MngeuCLpJeyTqAz3%2B4pT3hVQ%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
