New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
@Jack : yes I read those posts, but I didn't see that issue as a cause of my network downtime. I thought that described a separate issue (disk-related, "read-only"). If it's indeed the cause of my network downtime on node ukovz1 then I'll humbly shutup and wait for you guys to fix it
@serverian ouch. Kernel, maybe?
The VMs stop responding to network when the disk goes readonly. Therefore, downtime.
We have tried lots of different kernels, including regular stock CentOS kernels. Same thing happened on all of them.
Ah, enlightenment! OK I'll shutup now and wait
I am curious, talking about bugs... As openvz VPS have "root", can they flash the bios?
No
And to clarify, these things continued to happen when there was no VPS on it.
@serverian is haunted. Only explanation.
Well, I don't know what happened and since when @severian works for DotVPS. However, I'm quite happy with the cheapest UK-VPS I found so far. It's just sitting there acting like a VPN-server and I like it.
I know @Jack and I have not always had our positives. But Jack and Oktay really deserve a big up for the service they are providing at the moment. My main dedi is with Swiftway and I have been really a pain in the ass for complaining about random freezes and CPU keep posting warnings regarding temperatures. I have been looking for over weeks why the dedi was so terribly slow. It costed more than 3 days, including some reinstalls to figure out that the CPU temperature was the problem. Once we figured out what was wrong, Swiftway had to migrate the hardware to a new chassis and add extra fans.
Just an example of some node drama. I always took VPSes working for granted. Since I own a dedicated server and most of it is unmanaged, I really got more respect for systemadmins. It can be really frustrating if something doesn't work and you've tested anything.
However, regarding to this thread: I'm thinking of an software problem. I mean: if you rebuild everything from other parts, buy new nodes (with different hardware) and still the same thing happens, it must be something with the software. I think.
@serverian is not working for DotVPS though, he bought them over
Have you tried to move the RAID card to a different PCIe slot?
@severian, could be to do with power savings features/bug - we had to work around the issue earlier on a particular server.
So right now you have changed out 100% of the hardware - are you running similar PSUs between the various systems?
Do you have other servers at this location? If so are they affected? Or do you just have a single server?
Is the power source clean? Is it UPS'ed? I've seen issues like this out in India last year and it turned out to be a dirty power source.
It was an LSI-9271 combined with default / newer drivers. If we use an older version it'd work with warnings.
@Jack - so the PSUs were never replaced?
Which datacentre is this?
Are the PSUs dimensioned large enough? Have you tried running this on another power source?
If you have someone on-site pull it and run it from their office for a few days. See if the problem recurs. How much other equipment do you have in this rack? Is any of the rest of it showing any signs?
@MarkTurner nice catch. It may be the PSU OR the power lines
If I was a DotVPS customer I would appreciate the useful updates you're providing here. Keep up the good work Oktay and Jack. Hardware issues happen to all of us at some point.
For the record in case anyone is wondering, I have a VPS with Backupsy. While I don't have direct experience with the DotVPS brand as long as Oktay is in charge, you are in good hands
@Jack @serverian did you try pcie_aspm=off ?
Tests have been completed and the new servers seem stable. We have started to move people to the new server.
No, but as you see the disks are not going completely offline. We were still able to read the data off the raid array.
Everybody is moved to the new hardware. RFO sent. One month worth of service credit has been added.
Again, we deeply apologize for the trouble.
Just read it. Excellent email, thanks