We have a Ubuntu server with several HDD partitions that are then served to Windows machines using Samba. We just added a new drive and we have a big problem with it.
On day 1, I was able to write ~2TB of data without any problem. Now, when requesting files, the transfer starts normally then freeze, seemingly at random, after ~5sec. Stopping the transfer and restarting it makes it able to start again, but it freeze again after another ~5sec. I removed the partition from Samba and the same behavior is happening when downloading files from FileZilla.
When Samba was still on, I have seen that sometime the partition would crash and was remounted in Read-Only. That never happened without Samba using this partition
Similar issue I found online
I have read a lot of forum thread about dirty_background_ratio, dirty_ratio, etc, but I am unsure if I should touch these settings since we already have 2 other drives / partitions that are working just fine for a couple years now.
Information about our setup
All other drives are in RAID enclosures like this one: https://www.startech.com/en-ca/hdd/s3520bu33er. The new one that I have problems with uses a different one since the old one was OTS everywhere + it's been discontinued. Here's the newer model: https://www.startech.com/en-ca/hdd/s352bu313r.
This enclosure was recommended to us by Startech since it should be a 1:1 replacement of the old one.
I know hardware raid enclosure does not always have the best reputation, but these have been working great for us so far. The enclosure was first tested directly plugged in a Windows Machine for about a week without any problem before I wiped it to configure it in Linux, so I would tend to says that the issue is probably not with the enclosure or the drives.
All 3 are plugged-in using USB 3 and runs Raid 1 in there enclosure (but reports as a single drive the Linux). The only thing I could maybe imagine is different is that I plugged the cable of the new one in a different USB controller. The new box came with a USB-C cable so I used the USB-C port on our motherboard.
Could that be the issue? I think I also had weird issue when I was connected to the same controller as the other 2, not sure tho, it's been a while. Motherboard model in question is a ROG STRIX X370-F GAMING with a Ryzen 7 1700X if that's of any help.
Where I need help
I clearly have a problem somewhere in there. The fact that a file that freezes the transfer can be downloaded without any issue if I cancel and retry is weird to me. I'm unsure if this is a hardware of software issue.
