search

bblum / landslide-simics

the landslide codebase (old simics version)
11 stars 9 forks source link

In id wrapper make XCLOSE only warn, not panic, when close fails #172

Open bblum opened 9 years ago

bblum commented 9 years ago

removing log files early by hand because of disk quota is da real deal

bblum commented 9 years ago

appears to sometimes work sometimes crash

crash example:

[JOB 1] progress: 54569/98344 brs (55.487739%), ETA 1d 4h 5s (elapsed 19h 59m 40s)
Aborting -- time up!
[JOB 2] progress: 59483/195459 brs (30.432434%), ETA 11d 12h 44m 45s (elapsed 19h 59m 21s)
Aborting -- time up!
[JOB 3] progress: 54830/360291 brs (15.218247%), ETA 6d 17h 20m 3s (elapsed 19h 59m 31s)
Aborting -- time up!
[JOB 7] progress: 58067/203463 brs (28.539327%), ETA 2d 7h 46m 53s (elapsed 19h 58m 7s)
Aborting -- time up!
[JOB 5] progress: 54505/168892 brs (32.271996%), ETA 3d 22h 41m 19s (elapsed 19h 59m 0s)
Aborting -- time up!
Landslide pid 22136 exited with status 0
Landslide pid 24506 exited with status 0
Landslide pid 17380 exited with status 0
Landslide pid 12644 exited with status 0
Landslide pid 15000 exited with status 0
Assertion failed: '__ret == 0'
Assertion failed: '__ret == 0'
failed close fd 28
failed close fd 11
./landslide: line 25: 10311 Bus error               
(core dumped) ./$IDEXE -L $WRAPPER_LOG "$@"
bblum commented 9 years ago

when it does work, it prints:

Aborting -- time up!
Landslide pid 30292 exited with status 0
Assertion failed: '__ret == 0'
failed close fd 11
Error: Disk quota exceeded

and continues