Dne 04. 04. 23 v 23:50 Kevin Fenzi napsal(a):
I'm still unsure if that means we only do gp3 or what is the list of
ones we do after this change?
Depends on your use case. For Copr we use gp3 for all volumes but - copr-dist-git (5TB) and copr-be data volume where we
store dnf repositories (3*12TB). We do not need high IOPS. And Throughput 144/250MB (baseline/boost) is enough for our
use-case. And sc1 cost only $0.015/GB while gp3 cost $0.08/GB. For our copr-be volumes that makes $520 vs $2880 (per month).
To sum it up:
https://aws.amazon.com/ebs/volume-types/
* sc1 - cheapest volume, good enough if max throughput 250 MB is enough for you. $0.015/GB-month
* st1 - still magnetic, IOPS and throughput is two times bigger than sc1, but the price is 3times higher: $0.045 GB-month
* gp2 - does not have sense at all now
* gp3 - 3k-16k IOPS, 125MB-1000MB/s Throughput. $0.08 GB-month
* io1 - IMHO does not have sense now
* io2* - with $0.125/GB-month + $0.065/provisioned IOPS-month I never consider this, maybe for latency sensitive DB volumes
For rootfs in fedora-infra you want gp3 only.
But generally speaking for public cloud images you want to preserve magnetic volume as well. E.g. my personal webserver
runs of Fedora cloud image in AWS (under my private account) and because it has just few hits per hour, the st1 volume
for rootfs and data volume is just fine. And I have 6GB volume running for whole month for just 9 cents.
Miroslav
_______________________________________________
infrastructure mailing list -- infrastructure@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to infrastructure-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/infrastructure@xxxxxxxxxxxxxxxxxxxxxxx
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
