On Mar 19, 2016, at 11:31 PM, S.Sai Prashath wrote:
> I'm spawning a child process using spawn-command npm package, i do this when
> the node server starts and then for every request i read the query value and
> hit the running child process with stdin. The stdout that comes out of the
> child process is an event stream and i add a listener to read the data and
> then will be sending the data to sse-channel instance.
>
> Everything works the way it should for the first request, the issue arises
> for subsequent requests. When i console.log the data inside the stdout's on
> listener handler then i could see that the data is getting cloned. Meaning,
> same set of data is getting repeated twice for the second request and thrice
> for the third request etc... I was researching this issue and possible
> solution for this would be to create independent streams for each request so
> that this data cloning would be avoided in the child.stdout's on listener
> handler. But i'm not sure whether this could resolve the issue. Could some
> one please suggest a way to overcome this hurdle.
>
> Code looks like below,
>
>
>
> var spawnCommand = require('spawn-command');
> var cmd = 'path to the binary file';
>
> module
> .exports = function (app) {
>
>
> var child = spawnCommand(cmd);
>
> srvc
> = {
>
>
> get: function(req, res) {
>
> child
> .stdin.write('a json object in a format that is expected by binary' + '\n');
>
> child
> .stdout.on('data', function() {
>
> console
> .log(''+ data);
>
>
> });
>
>
> }
>
>
> }
> }
> As i mentioned earlier the child process would be spawned once when the node
> server starts and then the get() method would run for every request. So each
> time when a request runs, the stdin's would be sent to the binary and then
> the "on" listener would log the data in stdout. This works perfectly for the
> first request, but when a second request is made (with the first request
> still logging data) i could see that the data's for both first and second are
> getting cloned twice and for third request the data's for each request is
> getting cloned thrice and it increases exponentially.
>
> On the other hand if i spawn a child everytime inside the get() method i can
> overcome this problem but unfortunately i cannot do that because there'll be
> 100's of requests and i cannot afford 100's of binary instance eating up the
> memory.
I don't see how you can avoid spawning a new instance of the binary for each
request. Otherwise, how are you going to keep straight which output from a
single instance of the binary should be sent to which request?
If you cannot afford that many binaries running at once, use a queue or pool or
other similar system that would limit the number of processes that can be
spawned simultaneously. If too many binaries are already running, any
additional requests would have to wait until those binaries quit and new ones
can be started.
--
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/DF8A2A2D-5700-4A23-9F70-E709B2FC0AC7%40ryandesign.com.
For more options, visit https://groups.google.com/d/optout.
