MSBuild fatal errors when build triggered by timer

classic Classic list List threaded Threaded
10 messages Options
Reply | Threaded
Open this post in threaded view
|

MSBuild fatal errors when build triggered by timer

Michael Fawcett
Using Hudson 1.314.

If I "Build Now", my job succeeds with multiple MSBuild 2.0 tasks for
compiling C++ projects.  If it is triggered by the @daily timer, it
fails with the following MSBuild error messages:

"project specific file" - fatal error C1073: Internal error involving
incremental compilation(compiler file
'F:\SP\VCTOOLS\Compiler\CxxFE\sl\P1\C\p0io.c', line 865)

"project specific file" - fatal error C1090: PDB API call failed,
error code '23' : '(

and yes, the open parentheses is left dangling in the error message.

Searching the Internet for those error codes led me to some MSDN bug
reports that were closed "By Design" with no workarounds posted.  Any
help would be appreciated.

--Michael Fawcett

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

RE: MSBuild fatal errors when build triggered by timer

Horton, Anne B
I don't think MSBuild support is implemented until Visual Studio 2010.

-----Original Message-----
From: Michael Fawcett [mailto:[hidden email]]
Sent: Monday, August 10, 2009 2:17 PM
To: Hudson Users
Subject: MSBuild fatal errors when build triggered by timer

Using Hudson 1.314.

If I "Build Now", my job succeeds with multiple MSBuild 2.0 tasks for
compiling C++ projects.  If it is triggered by the @daily timer, it
fails with the following MSBuild error messages:

"project specific file" - fatal error C1073: Internal error involving
incremental compilation(compiler file
'F:\SP\VCTOOLS\Compiler\CxxFE\sl\P1\C\p0io.c', line 865)

"project specific file" - fatal error C1090: PDB API call failed,
error code '23' : '(

and yes, the open parentheses is left dangling in the error message.

Searching the Internet for those error codes led me to some MSDN bug
reports that were closed "By Design" with no workarounds posted.  Any
help would be appreciated.

--Michael Fawcett

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

Michael Fawcett
On Mon, Aug 10, 2009 at 3:02 PM, Horton, Anne B<[hidden email]> wrote:
> I don't think MSBuild support is implemented until Visual Studio 2010.

It spawns vcbuild.exe, so you're right, it's not native until 2010.
That's not the issue though.  It builds fine when I click Build Now.
It also builds fine using MSBuild locally on my machine.  The issue
only occurs when a timer triggers the build.

--Michael Fawcett

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

Marco Ambu-3
In reply to this post by Michael Fawcett
Michael Fawcett wrote:

> Using Hudson 1.314.
>
> If I "Build Now", my job succeeds with multiple MSBuild 2.0 tasks for
> compiling C++ projects.  If it is triggered by the @daily timer, it
> fails with the following MSBuild error messages:
>
> "project specific file" - fatal error C1073: Internal error involving
> incremental compilation(compiler file
> 'F:\SP\VCTOOLS\Compiler\CxxFE\sl\P1\C\p0io.c', line 865)
>
> "project specific file" - fatal error C1090: PDB API call failed,
> error code '23' : '(
>
> and yes, the open parentheses is left dangling in the error message.
>
> Searching the Internet for those error codes led me to some MSDN bug
> reports that were closed "By Design" with no workarounds posted.  Any
> help would be appreciated.
>
> --Michael Fawcett
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>  
Hi, I have the same problem (error C1090...) building c++ projects using
visual studio compiler called by ant scripts (not using MsBuild), but
unfortunately I have no solution.

I found a workaround: reducing the number of concurrent builds (the
number of executors) this problem is reduced (of course this is very bad
because we loose the build parallelism).
We have this kind of error only using hudson, but I cannot event think
how the build using visual studio command line tools can be "affected"
by hudson. However running several builds in parallel from dos prompt
does not generate this error.
Unfortunately this problem is completely random and "undebuggable" and
we have a lot of failed build with a waste of build time.
Microsoft states that this is a problem generated by a build either
using a network share or modifying the pdb folder.

--
Marco Ambu
<[hidden email]>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

gordin
In reply to this post by Michael Fawcett
Hi,
I run into the same error using a customized build system which uses Visual Studio's devenv.exe under the hood to compile our VisualStudio 2005 projects (VC++). I start two parallel builds (on different code base) every night. Always, the second job will fail with "Fatal error C1090: PDB API call failed, error code '23' : '(" in exactly the same second the first job finishes processing.
I analysed this thoroughly and can track the problem down to mspdbsrv.exe which is automatically spawned when building a VisualStudio project. All Visual Studio instances share one common pdb-server which shutdown itself after a idle period (standard is 10 minutes).
I assume that jenkins will clean up when a automatically started job finishes (like as described at http://wiki.jenkins-ci.org/display/JENKINS/Aborting+a+build). I checked mspbsrv.exe with ProcessExplorer and the process indeed has a variable JENKINS_COOKIE/HUDSON_COOKIE set in its environment if started implicitly by devenv.exe through jenkins. Don't know if this is evidence for my theory.

But, I can work around the problem by starting the pdb-server manually (not from within jenkins). Unfortunately, the program is not documented, but the important parameters are "-start -spawn -shutdowntime <seconds> -verbose" (see http://social.msdn.microsoft.com/Forums/en-US/vcgeneral/thread/b1d1bceb-06b6-47ef-a0ea-23ea752e0c4f/). Furthermore I need to configure a rather long idle time period because I do not know when the actual build starts (the jobs need to update a very large codebase from a cvsnt server, first). I don't know if there are any caveats when setting the shutdown time to large. As said earlier, the tool is undocumented, standard idle time is 10 minutes. I use a idle time of 5 hours. This works ok for me so far.

Long story short, while it is a handy feature to clean up after a build, the mechanism fails with Visual Studio's mspdbserv.exe which, when killed instead of gracyfully shut down after a idle period, will crash other builds.

I'm using Jenkins ver. 1.399

Any comments? Should I fill in a bug report for this issue?
--
Christoph Vogtländer
cek
Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

cek
gordin wrote
Any comments? Should I fill in a bug report for this issue?
I am seeing exactly the same issue and symptoms here.  If you don't submit a bug report, I will. :)
Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

gordin
On Thursday 10 March 2011 02:01:10 cek [via Jenkins] wrote:
> gordin wrote:
> > Any comments? Should I fill in a bug report for this issue?
>
> I am seeing exactly the same issue and symptoms here.  If you don't submit
> a bug report, I will. :)

I tried to get an account to be able to get access to JIRA, but I never got
the promised E-Mail. So, if you have an account please feel free to submit  
the bug report.

--
Christoph
Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

Erik Ramfelt
I know there has been problems in the past getting that signup email, have you tried logging in anyway?

Regards
//Erik

On Thu, Mar 10, 2011 at 02:34, Christoph Vogtländer <[hidden email]> wrote:
On Thursday 10 March 2011 02:01:10 cek [via Jenkins] wrote:
> gordin wrote:
> > Any comments? Should I fill in a bug report for this issue?
>
> I am seeing exactly the same issue and symptoms here.  If you don't submit
> a bug report, I will. :)

I tried to get an account to be able to get access to JIRA, but I never got
the promised E-Mail. So, if you have an account please feel free to submit
the bug report.

--
Christoph

Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

gordin
On Thursday 10 March 2011 09:55:21 Erik Ramfelt [via Jenkins] wrote:
> I know there has been problems in the past getting that signup email, have
> you tried logging in anyway?

I do not have a password. I never got the initial mail. I also tried "Reset
your password" without luck.

--
Christoph
Reply | Threaded
Open this post in threaded view
|

Re: MSBuild fatal errors when build triggered by timer

gordin
In reply to this post by cek
see http://issues.jenkins-ci.org/browse/JENKINS-9104
--
Christoph

On Thursday 10 March 2011 02:01:10 cek wrote:

> gordin wrote:
> > Any comments? Should I fill in a bug report for this issue?
>
> I am seeing exactly the same issue and symptoms here.  If you don't submit
> a bug report, I will. :)
>
>
> --
> View this message in context:
> http://jenkins.361315.n4.nabble.com/MSBuild-fatal-errors-when-build-trigge
> red-by-timer-tp385181p3344966.html Sent from the Jenkins users mailing list
> archive at Nabble.com.