Start Free Trial
Synology RS2414rp+ SLOW READ PERFORMANCE.
I setup 2x SSD's samsung 1TB on raid1 that is gonna be used for a few VMs for the next 2 months.
i setup raid 1 with BTRFS. Setup a single lun with advanced features. connected to it with ms iscsi initiator.
Everything works. I see the drive, i initialize it and format it with NTFS.
The Rackstation is connected with 1 gb cable to a gigabit switch. Jumbo frames 9000 enabled.
Big files transfer really really fast. back and forward. (maxed out 125 mb/s)
What i have an issue is, when i run atto benchmark on it, all the READS are maxed out at jokingly 5 to 11 mb/s??
The write speeds are doing better but not by much until i reach bigger files.
I am monitoring cpu and network usage on the GUI but there is absolutely no load / stress on the machine.
Supposedly this rackstation can do better than that?.
Here is what i have tried:
- Enable / disable jumbo frames = no change
- Trying out EXT4 instead of BTRFS = Gave around 20% boost but reads still stuck at 11 mb/s
- Trying different allocation sizes with NTFS (4k to 64k all the way) = small difference
- Trying ReFS = No difference
- Tried SMB3 mapped drive = no difference.
- Turning off all unnessercary services = no difference.
- Tried 3 different machines with different OS (server 2012R2, server 2016 and server 2019) = No difference
- Tried the raid "Sync faster" options = gave better results with normal operation but benchmark still shows bad.
- Tried directly cabling servers to the SAN with ip (no gateway). Same issue.
- Tried adding 4 TB WD reds. Same bad performance on read.
There is no MPIO setup.. just single Gigabit connection with iscsi.
Any device on the switch can transfer gigabit speeds with no issue.
Please help.. Getting desperate.
8/22/2022 - Mon
The 11 mb/s cap looks suspiciously like maybe your routing table has set a 10mb/s interface card with a preferred network metric, where traffic sometimes flows over a slower pathway.
Setup iperf3 + test your raw IP speed, then move on to testing your DNS lookup correctness (same IP from all NS servers), then move on to checking your routing table has correct metrics to prefer fastest network at all times.
I just ran iperf between the RP2414rp+ and local server. See attached for result.
It looks fairly normal for a gigabit connection.
Works perfectly on ubuntu!
So why is server 2016 / 2019 not doing well? How do we fix it on windows?
Experts Exchange has (a) saved my job multiple times, (b) saved me hours, days, and even weeks of work, and often (c) makes me look like a superhero! This place is MAGIC!
ASKER CERTIFIED SOLUTION
THIS SOLUTION ONLY AVAILABLE TO MEMBERS.
View this solution by signing up for a free trial.
Members can start a 7-Day free trial and enjoy unlimited access to the platform.
See Pricing Options
Start Free Trial
GET A PERSONALIZED SOLUTION
Ask your own question & get feedback from real experts
Find out why thousands trust the EE community with their toughest problems.
Plans and Pricing
Certified Expert Program
© 1996-2023 Experts Exchange, LLC. All rights reserved. Covered by US Patent