From: "Daniel Jitnah" <[email protected]>
> 1.Linux (Debian Wheezy) running on a cloud server
> 2.Everything working fine for weeks/months
> 3.Last nite, for some reason all disk writes to server stops
> 4.Even root cannot write to anywhere on the disk, and plenty of space
> left on disk (~80%)
>       All file permissions look OK.
>
> 5.After reboot all return to normal.(ie problem is fixed)
>
> My questions:
>
> Is it fair to assume that the problem was caused by something that the
> cloud provider has done? Or could it be something on server OS side?
>
> What can cause this?
>
> (I am thinking the virtual disk hosting the VM has become readonly
> somehow, but how? )

I had this in the past, with a VirtualBox VM (Ubuntu Server) on FreeBSD ZFS.

The performance of the ZFS became very bad - resulting in disk write
timeouts in the Ubuntu VM.

The Ubuntu server assumed a faulty disk and remounted the disk as
read-only to prevent further damage.

Although you are very unlikely to have FreeBSD/ZFS underneath there are
other ways to "produce" high I/O latency.

If so you should find these disk I/O related errors in your kernel logs.

I am afraid there is not much you can do (if you do not run the host
system). It only needs a short period with this problem and your system
becomes unresponsive.

Monitoring.. I did not do that for the read/only mount but all services
running on that VM turned red on the Icinga webpage.

Regards
Peter

_______________________________________________
luv-main mailing list
[email protected]
http://lists.luv.asn.au/listinfo/luv-main

Reply via email to