G
Guest
I have written a Windows Service which uses a FileWatcher to process file
dropped in a directory specified in a config file. If directory does not
exist it creates it in the constructor.
This works fine on my machine. We have created a new domain user specially
for the service that is added to the local adminstrators group and so should
have all the permissions needed to do anything related to files and messaging.
The service works fine on the machine on which it was developed. I created
an installer and rolled it out to a machine where our testers can access it.
From there it installs successfully on a test machine for which we have an
image to fall back to that is bare of any installations except the framework
(v1.1)
The machine has the service domain user added to the local administrators
group.
It works. Service installs using installutil.exe, is visible in services in
ComputerManagement snap-in. Service starts and does all that is supposed to.
Stop, uninstall, etc re-install and do it all again any number of times over.
At a particular point in time, the service ceases to start but only on the
test machine. It installs fine, but on attempting to start returns a 1053
error.
Check user, is OK, password not expired etc. Nothing appears wrong. Try
everything that we can think of without success. Wasted a day of 2 peoples
time trying to resolve this.
Finally roll operation system back to stored image and everything works
again.
Why?
How can we be sure of the ability of the service to do what we need it to do
without worrying if it will start at all? Or simply to not start at some
stage!
The only thing that might make any sense is that rollouts by system
administrators to tighten down security may have affected the ability to use
the user account, but again the question is how and why?
Following any of the articles related to Error 1053 would lead one to think
that SP1 for version 1.1 of the framework was not installed. Installed it.
Uninstalled service, reboot, reinstall, run installutil.exe from
....\framework\v1.1.4322, installs OK and creates directories as per
constructor. Service still will not start.
Why should it be necessary to roll back to a clean operating system
installation? We have no confidence that if we roll this out to a server for
production that it will install and start as required
This is being thrown back to me as a developer but I think the issue is
operating system stability.
Thoughts, comments, anything that might help would be very welcome right now!
cheers
dropped in a directory specified in a config file. If directory does not
exist it creates it in the constructor.
This works fine on my machine. We have created a new domain user specially
for the service that is added to the local adminstrators group and so should
have all the permissions needed to do anything related to files and messaging.
The service works fine on the machine on which it was developed. I created
an installer and rolled it out to a machine where our testers can access it.
From there it installs successfully on a test machine for which we have an
image to fall back to that is bare of any installations except the framework
(v1.1)
The machine has the service domain user added to the local administrators
group.
It works. Service installs using installutil.exe, is visible in services in
ComputerManagement snap-in. Service starts and does all that is supposed to.
Stop, uninstall, etc re-install and do it all again any number of times over.
At a particular point in time, the service ceases to start but only on the
test machine. It installs fine, but on attempting to start returns a 1053
error.
Check user, is OK, password not expired etc. Nothing appears wrong. Try
everything that we can think of without success. Wasted a day of 2 peoples
time trying to resolve this.
Finally roll operation system back to stored image and everything works
again.
Why?
How can we be sure of the ability of the service to do what we need it to do
without worrying if it will start at all? Or simply to not start at some
stage!
The only thing that might make any sense is that rollouts by system
administrators to tighten down security may have affected the ability to use
the user account, but again the question is how and why?
Following any of the articles related to Error 1053 would lead one to think
that SP1 for version 1.1 of the framework was not installed. Installed it.
Uninstalled service, reboot, reinstall, run installutil.exe from
....\framework\v1.1.4322, installs OK and creates directories as per
constructor. Service still will not start.
Why should it be necessary to roll back to a clean operating system
installation? We have no confidence that if we roll this out to a server for
production that it will install and start as required
This is being thrown back to me as a developer but I think the issue is
operating system stability.
Thoughts, comments, anything that might help would be very welcome right now!
cheers