ECONNRESET

Hi There,

First, as always I like to say thank you for your hard work on this library.

Recently I wrote a syncer script using jsforce, and I can’t help but notice when executing a very large query on a very large table the connection gets reset multiple times.

Now I have built in recovery procedure into the script to handle the problem, but my Project Manager asked me if there was anything I could do to increase the timeout time on my end.

My very simple log file looks like this:

starting another job: Task
begin working on job Task
updated job Task status {running:true}
beginSync
{ [Error: socket hang up] code: 'ECONNRESET' }
gotResults
salesforce error: { [Error: socket hang up] code: 'ECONNRESET' }
beginSync
{ [Error: socket hang up] code: 'ECONNRESET' }
gotResults
salesforce error: { [Error: socket hang up] code: 'ECONNRESET' }
beginSync
{ [Error: socket hang up] code: 'ECONNRESET' }
gotResults
salesforce error: { [Error: socket hang up] code: 'ECONNRESET' }
beginSync
gotResults
updating job Task lastDate:2014-11-14T00:37:00.000+0000
continueSync
gotResults
updating job Task lastDate:2014-11-19T19:31:30.000+0000
continueSync
...

After three hangups, everything begins to work correctly. This has been a consistent pattern with one table and query.

Please do let me know if there is some setting on my end I can control to extend how long the connection is open for.

Thank you!

[seanfuture]

We submit a lot of batch jobs to Salesforce on a daily basis and I can tell you 1st hand that we see this error regularly. We process about 100-200 batch jobs a day. It’s become much more frequent in the past 3-4 months. I agree with @tejasmanohar in that we get these most commonly when trying to close jobs - As though Salesforce simply can’t get the job done at times and so the API call will timeout. All-in-all their API is extremely unreliable.

Currently forced to issue retries and/or alternative designs that limit the amount of jobs needed just to attempt to avoid the unreliable Salesforce job system. Suggestions welcome however short of opening a support ticket with Salesforce I believe this is out of jsforce’s control. Please leave this ticket open for now in case a creative solution presents itself.

[ArcusSimonCull]

Encountered this issue when sending multiple queries through a Promise.all(). Interestingly the first retry hit an issue where a query returned 0 results (not possible with ‘SELECT Id, DeveloperName FROM RecordType’). The next retry succeeded with no issue. If this is a timeout issue, is there a way to define the timeout for the http.request() to allow more time for SF to complete the request at “busy” times?