Unsolved
This post is more than 5 years old
1 Message
0
655
November 18th, 2019 08:00
Differing write speeds comparing 2 FX2s
Hi All. We currently run 2 FX2s for our Hyper-V VM environment. On the FX2 we purchased two years ago, everything seems fine. But on the second (newer-purchased about a year ago) one, we are seeing a severe decrease in write speeds. What is giving us a headache is that the write speeds are happening on both compute nodes in the newer FX2. The chassis are populated by two compute and two storage nodes. When we ordered the new FX2, we requested that the configuration be as similar as possible to the one we purchased a year prior. While there are some differences, they don't seem to be enough that we would see such a severe difference. To the point where a new replica database never fully replicates 1.5TB of data in a day.
The other strange thing is that large file transfers appear to be fine, but when lots of files are being written we see the slowness. As shown below, DBR02 is on the older FX2 and has normal speeds, but VM18 is on the newest FX2 and is significantly slower.
Again, moving VMs onto the "old" FX2, they run fine. Move the same VMs to either of the computer nodes on the new FX2, and we consistently get the slow speeds.
We're stumped at this point because there are no Event errors on either Hyper-V hosts or in the VMs themselves. I'm thinking there's a hardware issue with the chassis, but I'm not sure how to run hardware tests on it without taking down the entire host. Does anyone have any suggestions that may help us find a solution? Both FX2s are plugged into two different Dell 4148T-ON switches configured for redundancy. Network tests imply that the issue is not network related.
arcit@dbr02:~$ dd if=/dev/zero of=/tmp/test1.img bs=1G count=1 oflag=dsync 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 5.15987 s, 208 MB/s 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 5.15409 s, 208 MB/s 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 5.2012 s, 206 MB/s arcit@vm18:~$ dd if=/dev/zero of=/tmp/test1.img bs=1G count=1 oflag=dsync 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.68147 s, 639 MB/s 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.6963 s, 633 MB/s 1073741824 bytes (1.1 GB, 1.0 GiB) copied, 1.69163 s, 635 MB/s arcit@dbr02:~$ dd if=/dev/zero of=/tmp/test2.img bs=512 count=1000 oflag=dsync 512000 bytes (512 kB, 500 KiB) copied, 1.84467 s, 278 kB/s 512000 bytes (512 kB, 500 KiB) copied, 1.62383 s, 315 kB/s 512000 bytes (512 kB, 500 KiB) copied, 1.28388 s, 399 kB/s arcit@vm18:~$ dd if=/dev/zero of=/tmp/test2.img bs=512 count=1000 oflag=dsync 512000 bytes (512 kB, 500 KiB) copied, 7.7669 s, 65.9 kB/s 512000 bytes (512 kB, 500 KiB) copied, 7.7007 s, 66.5 kB/s 512000 bytes (512 kB, 500 KiB) copied, 7.63801 s, 67.0 kB/s
We greatly appreciate any insight!


