Re:EC2 Instance Update – C5 Instances with Local NVMe Storage
by lathiat on Friday May 18, 2018 @06:11AM
(#17100187)
According to the bug here:
https://bugs.launchpad.net/ubuntu/bionic/+source/linux/+bug/...
It was actually filed about EBS block disks using NVME (on these new instances, there is a hardware card that presents network EBS volumes as a PCI-E NVME device). In certain failure cases since this is a network block store, they can fail for some period of time exceeding this timeout.
The idea of this change is to ensure once they come back the machine is left in a usable state.
Of course, I would not expect to see this on Local NVME disks which is what they announced - that you can now get such instances with local disk as well as EBS.
Re:EC2 Instance Update – C5 Instances with Local NVMe Storage
by ohre1Eda on Friday May 18, 2018 @07:50AM
(#17100648)
Ah yeah, the article was about local NVMe, so the concern is probably not relevant here.
Re:EC2 Instance Update – C5 Instances with Local NVMe Storage
by lathiat on Friday May 18, 2018 @11:23AM
(#17102185)
Unrelated: love your pwgen username
Re:EC2 Instance Update – C5 Instances with Local NVMe Storage
by benmmurphy on Friday May 18, 2018 @07:20AM
(#17100485)
EBS probably has really terrible tail latencies.
I can imagine some crazy random access, small-record, vectored IO from large thread pools. But that's not exactly common because most software that is IO-heavy tries really hard to avoid these things.