Why my EC2 instance frozen?

Question

My t2.micro instance becomes inaccessible when I run two processes, each loads about 112 MiB of data (plus some metadata overhead).  The Instance state is still "Running".  But I cannot ssh to it, neither can I send requests to it anymore.  I can reboot my instance and ssh to it again, but running the two processes with those data will freeze it again.

I suspect that the two processes consume all memory in my instance.  But the behavior is strange---I would expect the two processes to OOM and crash, rather than the instance freezes.  In addition, the last report of `free -m` command (which is run every second until the terminal is disconnected from the instance) shows memory remains:
               total        used        free      shared  buff/cache   available
Mem:             949         377         377           3         194         438
Swap:              0           0           0

What could cause this problem?

Answer

I'd suggest you to setup CloudWatch monitoring to observe CPU, memory  metrics before launching those two processes.

Once instance freezes, check cloudwatch metric and see where the resources limits are being hit. You might need to consider an instance of higher configuration or two instances of t2.micro and let only one process run on each.

See [Cloudwatch Monitoring](https://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/WhatIsCloudWatch.html) for more details.

Answer

This usually happens when the instances memory is maxed out.

I have seen this several times and the fix is to upgrade to an instance type with more RAM.

Why my EC2 instance frozen?

相關內容