NetBSD-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

NetBSD on Linode kvm problems



I've been running a NetBSD VPS with Linode for a number of years now.
I started with 6.0 back when it was Xen based, and converted over to
KVM when they announced they were moving us over to the different
hypervisor.

It took a bit of twiddling with the settings (use full vitalization,
not para-vitalization, turn off all the OS 'helpers', etc), renaming
devices (/dev/vdX->/dev/wdXn, new NIC names, etc.), but it was working.

A few weeks ago, Linode patched their systems for Meltdown, and that's
where things have gone south for me.  The VM got slow, like, super slow.
Doing things like reading mail, over imap or directly, took forever to
open mailboxes, move to new messages, etc.  My mosh session would
repeatedly tell me how it had lost connection briefly and come back.

I then noticed that my clock was way off.  After about a week of
running, the clock was 4 days(!) behind, even with ntpd running.
Killing ntpd and running ntpdate to bring it back current would take
minutes to return, and it would quickly fall behind again.

I upgraded from 6.1.4 to 7.1.1, hoping the maybe the newer kernel would
handle things better, but again, after running a few days, things got
slow again, my clock was way off, etc.  Attempting a clean reboot took
over 30 minutes before I gave up and pulled the virtual plug and did a
'hard' reset.

I'm running the 'NetBSD 7.1.1 (GENERIC.201712222334Z) amd64' kernel from
the install set, but I'd be willing to compile a new one if there are
some kernel options I can tweak to make it work better.

Does anyone have any suggestions for things I should look at/fiddle with?

Thanks.

--
Michael Parson
Pflugerville, TX
KF5LGQ


Home | Main Index | Thread Index | Old Index