[andy@bitfolk.com: Re: Host "hen" crashed again (Was: Re: Host "hen" unexpectedly rebooted 2018-11-26 22:24)]
21 Dec
2018
21 Dec
'18
6:43 p.m.
Oops, I've just noticed that this only went to users@, not announce@. Sorry about the
extra copy some of you are now getting. :(
Cheers,
Andy
----- Forwarded message from Andy Smith <andy(a)bitfolk.com> -----
Date: Fri, 21 Dec 2018 14:49:34 +0000
From: Andy Smith <andy(a)bitfolk.com>
To: users(a)lists.bitfolk.com
Subject: Re: [bitfolk] Host "hen" crashed again (Was: Re: Host "hen"
unexpectedly rebooted 2018-11-26 22:24)
User-Agent: Mutt/1.5.23 (2014-03-12)
Reply-To: users(a)lists.bitfolk.com
Hi,
This happened again a few minutes ago.
IPv6 failover appears to have happened correctly due to some fixes
since last time - hence no v6 outage until manual intervention.
hen has now booted with some settings which it is hoped will avoid
the problem in future.
I'm not in front of a computer at the moment so later today I'll
review what was logged.
Apologies for the ongoing hassle.
Cheers,
Andy
_______________________________________________
announce mailing list
announce(a)lists.bitfolk.com
https://lists.bitfolk.com/mailman/listinfo/announce
21 Dec
21 Dec
7:20 p.m.
Hello,
On Fri, Dec 21, 2018 at 06:43:26PM +0000, Andy Smith wrote:
I'm not in front of a computer at the moment so
later today I'll
review what was logged.
The hypervisor appears to have crashed servicing the same guest as
it was servicing last time it crashed. It is very unlikely to be
malicious, just some strange interaction with that guest.
I have contacted that customer to let them know I'll move
segregating their VPS onto test hardware in the next 6 hours.
"hen" is now running with configuration options which are thought to
have a good chance to avoid the area of code the bug is thought to
lurk in, so the test host will not use those options and we'll see
if this guest crashes it.
So, in summary, we've learned some useful things and configuration
is now set that it's hoped will help avoid this problem on "hen"
again, but we aren't 100% sure yet.
Cheers,
Andy
--
https://bitfolk.com/ -- No-nonsense VPS hosting
_______________________________________________
announce mailing list
announce(a)lists.bitfolk.com
https://lists.bitfolk.com/mailman/listinfo/announce
1 Jan
1 Jan
7:52 p.m.
Hello,
On Fri, Dec 21, 2018 at 07:20:08PM +0000, Andy Smith wrote:
I have contacted that customer to let them know
I'll [be]
segregating their VPS onto test hardware in the next 6 hours.
10 days later, this guest crashed the test server. I wouldn't
normally be happy that a server has crashed but at least the problem
can be replicated now with that one guest¹.
The test server has now been booted with the configuration that was
suggested may avoid the problem, so it's now a matter of waiting to
see if that is the case. I think the longest we've gone without
issue has been 14 days.
Cheers,
Andy
¹ I should stress again that I don't believe the guest is doing
anything malicious, it's just doing something to tickle a bug that
nobody else seems to be doing. I suspect that potentially any
guest on any server could do this at any time.
--
https://bitfolk.com/ -- No-nonsense VPS hosting
_______________________________________________
announce mailing list
announce(a)lists.bitfolk.com
https://lists.bitfolk.com/mailman/listinfo/announce
1935
days inactive
1946
days old
2 comments
1 participants
participants (1)
-
Andy Smith