[sheepdog-users] Snapshot and Cache stress test

[Valerio Pachera]

2013/12/17 Liu Yuan <namei.unix at gmail.com>

> This means some IOs are not returned yet. I think probably you can log in
> the
> VM as a another user to test if the VM is really deady or just some IO
> tasks
> are hung
>

I've been waiting several minutes.
The prompt is frozen.
I didn't try to change virtual console because I couldn't by vnc.
Anyway, I find it strange that the guest doesn't die by kill -15.
In other words something fatal happens to the qemu process, even if the
guest is not fully frozen.

I think there's something overlapping.
This seems mostly to happen after I removed 8G of data, snapshot, then
started adding new data and snapshot.

After removing 8G of data I even run 'sync' and it was instantaneous.
Maybe sheep wasn't that fast.

Sorry if I say inaccurate things, I'm just trying to guess were the issue
could be.

(I think in normal working condition it's difficult to step in it, but it
may happens if we use the snapshots a lot in a heavy I/O environment).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.wpkg.org/pipermail/sheepdog-users/attachments/20131217/d2ace2d5/attachment-0005.html>