I just migrated my application from Meteor 1.2 to Meteor 1.8.
Now my server goes down randomly, sometimes after 3 hours, and sometimes after 6 hours, its totally random.
When the server is down, my mobile apps or web app stops working, so I restart my server and after that it remains slow for a minute or two, and it became faster again.
When the server is down, at that time the node uses 105% of CPU usage.
Have you allowed it to run in developer mode for that amount of time to see if you can catch the bug?
Also, I set up a special collection called error logs, that I didn’t write any errors to go whenever they are caught in the code. It’s not ideal when running in production mode but I set up a few extra parameters that I sent to the log to let me know which file method in line number the error happened on at least according to where the error is thrown.
Only thing I see is Babel runtime twice, but probably not that.
So, I want to understand, when you say the server stops, do you mean it crashes, or it’s still running in the terminal, but the pages stop responding, or become really slow?
Nice rig, so memory definitely shouldn’t be the issue. Guessing that out of memory error was a 1-off. I wish I had more ideas. After updating to 1.8, did you run `meteor update —all-packages as well?
@bmcgonag1 Few years back I integrated Kadira plugin, which would help check the performance statistics and crashes etc, since it has been discontinued, so is there any alternate for that?
But what @minhna says might be the issue, didnt saw that. But how are you even able to run meteor 1.8 with node 0.10 when using meteor --port NNNN. That should use the node version meteor comes with.
I am also getting this error time to time while moving in paginations:
Executor error during find command :: caused by :: Sort operation used more than the maximum 33554432 bytes of RAM. Add an index, or specify a smaller limit.