search

git-for-windows / git

A fork of Git containing Windows-specific patches.
http://gitforwindows.org/
Other
8.38k stars 2.54k forks source link

Investigate CI failures due to "error sending signal 11" #5222

Open dscho opened 1 week ago

dscho commented 1 week ago

Every once in a while, a win test or win+VS test job in our CI runs fails with a very obscure "Parse errors: No plan found in TAP output" bread-crumb left behind in the prove output. Example:

  [23:56:45] unit-tests/bin/t-strbuf.exe ..................... ok        1 s ( 0.00 usr  0.02 sys +  0.03 cusr  0.17 csys =  0.21 CPU)
  [23:56:45] t9135-git-svn-moved-branch-empty-file.sh ........ skipped: skipping git svn tests, NO_SVN_TESTS defined
  [23:56:45] unit-tests/bin/t-strcmp-offset.exe .............. ok        1 s ( 0.00 usr  0.02 sys +  0.54 cusr  1.31 csys =  1.87 CPU)
  [23:56:45] unit-tests/bin/t-trailer.exe .................... ok        1 s ( 0.00 usr  0.00 sys +  0.51 cusr  1.29 csys =  1.80 CPU)
  [23:56:45] unit-tests/bin/t-urlmatch-normalization.exe ..... ok        1 s ( 0.00 usr  0.00 sys +  0.53 cusr  1.27 csys =  1.80 CPU)
  [23:56:45] unit-tests/bin/unit-tests.exe ................... ok       <1 s ( 0.00 usr  0.00 sys +  0.54 cusr  1.24 csys =  1.79 CPU)
  [23:56:45] t0071-sort.sh ................................... ok        4 s ( 0.00 usr  0.08 sys +  3.74 cusr  9.97 csys = 13.79 CPU)
  [23:56:48] unit-tests/bin/t-reftable-stack.exe ............. ok        4 s ( 0.00 usr  0.05 sys +  1.00 cusr  2.67 csys =  3.72 CPU)
  [23:56:52] t0302-credential-store.sh ....................... ok      108 s ( 0.09 usr  0.30 sys + 72.16 cusr 177.20 csys = 249.76 CPU)
  [23:56:55] t3906-stash-submodule.sh ........................ ok       37 s ( 0.02 usr  0.16 sys + 28.74 cusr 72.28 csys = 101.20 CPU)
  [23:56:58] t1300-config.sh ................................. ok      302 s ( 0.24 usr  0.58 sys + 146.59 cusr 356.84 csys = 504.24 CPU)
  [23:56:58]

  Test Summary Report
  -------------------
  t7513-interpret-trailers.sh                   (Wstat: 256 (exited 1) Tests: 24 Failed: 0)
    Non-zero exit status: 1
    Parse errors: No plan found in TAP output
  Files=124, Tests=2693, 302 wallclock secs ( 0.25 usr  0.61 sys + 146.59 cusr 356.84 csys = 504.29 CPU)
  Result: FAIL
  make: *** [Makefile:73: prove] Error 1
  ++ cat exit.status
  + res=2
  + rm exit.status
  + end_group 'Run tests'
  + test -n t
  + set +x
=== Failed test: t7513-interpret-trailers ===
The full logs are in the 'print test failures' step below.
See also the 'failed-tests-*' artifacts attached to this run.
Error: Process completed with exit code 1.

Looking at those test failures shows some two dozen lines with the error mentioned in this ticket's title. Example:

      0 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
    962 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
./test-lib.sh: fork: retry: Resource temporarily unavailable
1554205 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
2566469 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
3112806 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
4122000 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
./test-lib.sh: fork: retry: Resource temporarily unavailable
4666365 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
6228598 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
7694614 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
7786764 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
9253621 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
./test-lib.sh: fork: retry: Resource temporarily unavailable
9346422 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
10904208 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
12460296 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
14020585 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
14818246 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
15582141 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
16373295 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
./test-lib.sh: fork: retry: Resource temporarily unavailable
17136204 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
18694786 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
20236375 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
21798608 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
23354580 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
24908795 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
25936958 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
26465152 [sig] sh 529 sig_send: error sending signal 11, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
27493410 [main] sh 529 sig_send: error sending signal -72, pid 529, pipe handle 0x14C, nb 0, packsize 176, Win32 error 0
./test-lib.sh: fork: Resource temporarily unavailable
FATAL: Unexpected exit with code 254

This error message has been mentioned on the Cygwin mailing list in 2022, seemingly without any traction seeing as the proposed patch never made it into Cygwin, at least I cannot find it, even on their master branch.

The same error message was mentioned again in July '24, this time on the cygwin-developers mailing list, again without any traction.

Maybe we should integrate the proposed work-around into the MSYS2 runtime and see whether it improves things?

rimrul commented 1 week ago

This error message has been mentioned on the Cygwin mailing list in 2022, seemingly without any traction seeing as the proposed patch never made it into Cygwin, at least I cannot find it, even on their master branch.

Looking at this reply I think it became this commit in 3.3.

dscho commented 1 week ago

This error message has been mentioned on the Cygwin mailing list in 2022, seemingly without any traction seeing as the proposed patch never made it into Cygwin, at least I cannot find it, even on their master branch.

Looking at this reply I think it became this commit in 3.3.

Good find, I had overlooked that email!

But then I am really at a loss in which direction to look for a hint at a solution...