I just started a new account here to see if anyone could help me with a problem I am having.
I have an OpenVZ VPS node with around 50 users on it, but the damn thing keeps going into read-only mode every few days and I have NO idea why this would happen. I have requested sys-admins and other techs to take a look and no one seems to know the issue. This is my hardware setup --
It is running the latest OpenVZ Kernel so I have no idea why the panel keeps going into read-only. The RAID card has been replaced once and all the drives tested out healthy. All hardware and software are updated to the latest drivers and firmware as well.
Does anyone have any ideas or experiences with the a similar issue? The load is low and everything is fast as hell, but this is killing the stability I was hoping for.
What you first need to do is get a better idea why its going read only. Check dmesg for an error: is it a file system error (possibly correctable with fsck) or some type of hardware error?
Next check the hardware. At minumum verify with smartcheck all the drives are in good health, and have no bad (uncorrectable/pending sectors).
If smartcheck checks out and it is a file system error, then an FSCK may just needed. Otherwise you need a more involved hardware test. I have seen read only errors from anything from bad power supply to RAM. A memtest can clear ram issues, and your DC should have a simple power supply tester.
If ram and the PSU pass, then you'll need to swap the hardware like the raid card, motherboard and/or CPU.
Yeah, in the past we have found RO issues tend to usually be random hardware related problems. Start with the simple thing, smart checks on the drives, memtest on the memory, replace the power supply, you can rule out the raid card since its very unlikely 2 cards are bad for you, then check to make sure your raid card is up to date with the latest firmware/bios as well as your MB. And lastly if you still have the issue swap the MOBO and CPU.