We have found that we can place 18 journals on the Intel 3700 PCI-e devices comfortably, We also tried it with fio adding more jobs to ensure that performance did not drop off (via Sebastian Han’s tests described at https://www.sebastien-han.fr/blog/2014/10/10/ceph-how-to-test-if-your-ssd-is-suitable-as-a-journal-device/) so 12 should be no problem – only gotcha is if the NVMe dies . . .
From: ceph-users [mailto:[email protected]] On Behalf Of Massimiliano Cuttini Sent: Friday, June 23, 2017 9:35 AM To: Ashley Merrick Cc: [email protected] Subject: Re: [ceph-users] Squeezing Performance of CEPH Hi Ashley, You could move your Journal to another SSD this would remove the double write. If I move the journal to another SSD, I will loss an available OSD, so this is likely to say improve of x2 and then decrease of x½ ... this should not improve performance in any case on a full SSD disks system. Ideally you’d want one or two PCIe NVME in the servers for the Journal. This seems a really good Idea, but image that I have only 2 slots for PCIe and 12 SSD disks. I image that it's will not be possible place 12 Journal on 2 PCIe NVME without loss performance..... or yes? Or if you can hold off a bit then bluestore, which removes the double write, however is still handy to move some of the services to a seperate disk. I hear that bluestore will remove double writing on journal (still not investigated), but I guess Luminous will be fully tested not before the end of the year. About the today system really don't know if moving on a separate disks will have some impact considering that this is a full SSD disks system. Even adding 2 PCIe NVME.... why should not use them as a OSD instead of journal solo?
_______________________________________________ ceph-users mailing list [email protected] http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
