I have very, very few users.
I was scrapping some data today but that's usually not a problem. I've attached a screenshot of my instance stats.
Everything has
normalized and the instance is still hanging. Do I have any hope here?
Asked
Active
Viewed 252 times
0
-
Which OS are you using ? – user9517 May 25 '11 at 20:39
-
Using Ubuntu 10.4 – Ben May 25 '11 at 20:41
-
When it hangs can you still use ssh to access it ? If you can what does top have to say ? – user9517 May 25 '11 at 20:47
-
I can't ssh into the instance, which is clearly not a good sign – Ben May 25 '11 at 20:53
-
1Try restarting it, if that doesn't work, detach the EBS-block and start a new instance. Don't get attached to your EC2-instance :-) – Bart De Vos May 25 '11 at 20:58
-
If I restart it I'll lose all the temporary files right? Like my entire Lamp Stack? I'm deploying Django, I don't have an AMI setup and it's a massive pain to recreate the Django Environment – Ben May 25 '11 at 21:01
-
As it is right now, you have no disk activity and can't ssh to it. I don't think you can salvage anything it without restarting. – Holocryptic May 25 '11 at 21:39
-
It is a Christmas miracle. I rebooted her and she somehow or other came back to life. – Ben May 25 '11 at 22:33
1 Answers
3
Amazon EC2 instances should be thought of as temporary, we have about 100 running and about 10% die every week, we just kill them off and start again. Just create a new instance from the EBS image (You created an EBS image right? ;) ).
At first I thought this was a bit odd for so many instances to fail on such a regular basis but I contacted AWS support and they said pretty much what I said above. Infact I don't think I've ever had an instance running for longer than a month without it failing (Although we do hammer all our instances, they're on 100% CPU pretty-much all the time)
-
This is really a useful point of view. I felt like a tool for blowing up an instance without any real traffic. Now I know that it just happens sometimes. Thanks – Ben May 25 '11 at 22:34