Re: [Xen-devel] [xen-unstable test] 61521: regressions - FAIL

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 08.09.15 at 17:45, <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> Ian Jackson writes ("Re: [Xen-devel] [xen-unstable test] 61521: regressions - 
> FAIL"):
>> A more likely explanation is that the host sometimes fails to complete
>> its bootup within a reasonable time.
> 
> Actually, looking at a bit more of the osstest transcript:
> 
> 2015-09-08 09:14:32 Z guest debianhvm.guest.osstest 5a:36:0e:51:00:2a 22 
> link/ip/tcp: ok. (18s)
> 2015-09-08 09:14:33 Z executing ssh ... root@xxxxxxxxxxxx echo guest 
> debianhvm.guest.osstest: ok
> ssh: connect to host 172.16.145.6 port 22: Connection refused
> 
> The first message means that
>   nc -n -v -z -w $interval $ho->{Ip} $ho->{TcpCheckPort}
> succeeded.  The IP address is the one from DHCP and the port is 22
> as you see printed.  $interval is 5.
> 
> So what this means is that the osstest controller successfully made a
> tcp connection to the guest's port 22.  It then went on to the next
> check which is actually to try ssh'ing to the host.  But ssh got
> ECONNREFUSED.
> 
> This means that the guest was accepting connections on 172.16.145.6:22
> and then stopped doing so (perhaps only briefly).

That's what I was about to say in a reply to your earlier mail. To me
this still means the guest booted up successfully (or else it wouldn't
have accepted the TCP connection). I agree its "ssh server" service
is unavailable (or unreliable), which is bad, but which then again we
have no real handle to deal with without someone running into it
outside of osstest.

Whether to use this as justification for a force push I'm not sure
anyway.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel