In 'top', when the server starts, the memory shows...
Mem: 1781972k total, 368760k used,
The 'used' memory slowly climbs till it reaches the 'total'. It is at
this point that the cpu goes to 100% and the instance becomes
unresponsive. S Moser pointed out that this isn't an actual reading of
memory availablity and directed me to 'free'
I used 'free' refreshing at 3 second intervals and on the last record
before it locked up it showed
total used free shared buffers cached
Mem: 1740 1465 275 0 369 729
-/+ buffers/cache: 366 1374
Swap: 895 0 895
So there was plenty of 'available' memory at that point.
It seems to me though that while only a small portion of memory is
actually in use the system is unable to allocate the 'used' memory back
when it's needed?
As a test I launched a 64 bit version from ami-a0b993d4 (instance-store,
eu-west-1 official Ubuntu 8.04 LTS Hardy) on an m1.large. Installed the
same packages, attached the EBS volume with the same PHP based website
etc. It has now been up 24hrs without an issue.
Not sure if this contributes anything as it's a different ami (with
presumably different 64 bit versions of the packages) but I think it
eliminates anything in our PHP code
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/745108
Title:
EC2 instance CPU goes to 100% and locks
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
