cygwin sshd service terminated unexpectedly.

[Junaid Ali <al...@iit.edu>]

Hello,

Sometimes the reserved VM boots up and the Cygwin service fails to start
during post_load and the management node gives up after trying for 1800
seconds resulting in a failed reservation.



I logged into the VM and found that Cygwin failed to start with error
message in update_cygwin.log



The CYGWIN sshd service could not be started. A system error has occurred.
System error 1067 has occurred. The process terminated unexpectedly. Error
level: 2



If I manually start the service it starts just fine. I checked that time
that the VM received IP's from the DHCP server and it was about 1 minute
before the update_cygwin script ran, so the VM had network connectivity
before the service was attempting start (I am not sure if sshd will crash
if network connectivity is not available).



The following are the software version details:

Cygwin: 1.7.25 (0.270/5/3) CYGWIN_NT-6.1-WOW64

OpenSSH (on VM): 6.4

VCL: 2.2.1

OS (VM): Windows 7 SP1 (64-Bit)



The Cygwin sshd service has the following options set:

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\sshd\Parameters\Environment\CYGWIN
ntsec nodosfilewarning



Would it help to have Cygwin sshd service recovery options set so it
restarts the service after First and Second failure? Would this interfere
with the way the post_load scripts are setup?



Has anyone seen this behavior? Any suggestions?



Thanks.

Junaid Ali

Re: cygwin sshd service terminated unexpectedly.

[Mike Haudenschild <mi...@longsight.com>]

I experienced something similar with VCL 2.2.x and it was hard to
troubleshoot because it was intermittent.  If you're using vSphere as your
hypervisor, make sure that VMWare Tools is installed and working.  If
there's *nothing* consistent about the problem (e.g. the same VM, IP
address conflicts, etc.) I might suggest re-running the CygWin install
procedure from the VCL base-image documentation and re-capturing the image.
 Make sure you're running the Cygwin install as root, make sure that
Windows UAC is turned off, make sure no anti-virus software is flagging it,
and check the file permissions to make sure SYSTEM has access.  See if
Cygwin SSHD is writing out a log file of its own that would give you more
information.  You could certainly try setting the recovery options, but
that seems at best like a band-aid.


On Thu, Feb 27, 2014 at 12:41 PM, Junaid Ali <al...@iit.edu> wrote:

> Hello,
>
> Sometimes the reserved VM boots up and the Cygwin service fails to start
> during post_load and the management node gives up after trying for 1800
> seconds resulting in a failed reservation.
>
>
>
> I logged into the VM and found that Cygwin failed to start with error
> message in update_cygwin.log
>
>
>
> The CYGWIN sshd service could not be started. A system error has occurred.
> System error 1067 has occurred. The process terminated unexpectedly. Error
> level: 2
>
>
>
> If I manually start the service it starts just fine. I checked that time
> that the VM received IP's from the DHCP server and it was about 1 minute
> before the update_cygwin script ran, so the VM had network connectivity
> before the service was attempting start (I am not sure if sshd will crash
> if network connectivity is not available).
>
>
>
> The following are the software version details:
>
> Cygwin: 1.7.25 (0.270/5/3) CYGWIN_NT-6.1-WOW64
>
> OpenSSH (on VM): 6.4
>
> VCL: 2.2.1
>
> OS (VM): Windows 7 SP1 (64-Bit)
>
>
>
> The Cygwin sshd service has the following options set:
>
> HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\sshd\Parameters\Environment\CYGWIN
> ntsec nodosfilewarning
>
>
>
> Would it help to have Cygwin sshd service recovery options set so it
> restarts the service after First and Second failure? Would this interfere
> with the way the post_load scripts are setup?
>
>
>
> Has anyone seen this behavior? Any suggestions?
>
>
>
> Thanks.
>
> Junaid Ali
>
>
>