Communication Lost 3 hours into the print

edited December 2015 in Bug Reports
Hi,

I have upgraded to 0.70.0 and three or so hours into the print I got communication lost message. Luckily I was there to see it happen. The print did not resume so I had to kill it. After checking the log files for the print job and the server log I see that I have a lot of these errors in the server log file (these messages were not present in previous jobs for 0.70.0 or 0.65.0) -

2015-12-06 21:40:56: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:40:57: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:40:58: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:40:59: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:00: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:01: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:02: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:03: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:04: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:05: boost thread: failed in pthread_create: Resource temporarily unavailable

2015-12-06 21:41:07: boost thread: failed in pthread_create: Resource temporarily unavailable


They repeat every second as you can see. Could you please look into this. I do not know if this is related to new time-lapse video or me running an external command every layer. What I noticed in the previous version of RS was that there were a lot of <defunct> processes left after a print job. I guess they were the external commands leftovers. In the new version of RS I do not see these anymore.

For now I will turn off the time-lapse video and external commands to see if I can reprint my objects without a problem.

Would appreciate any kind of hint as to what I might be doing wrong. Or perhaps it is a genuine bug :-(


Regards,

yOOrek

Comments

  • I guess it is your external command every layer. To solve the zombie porblem we have to wait for process to finish, which we do in a threaded process created for this. So what I think happens is that you exceed the limit of running threads allowed at some point. Will check that assuming your function finish in time. So maybe threads need to be joined to free them also I never needed so far. Timelaps has only one thread all the time of printing, so nothing compared with your thread per layer:-)
  • I could reproduce the problem with a file that only called /usr/bin/w and waited 50 milliseconds. Crashed after a few seconds. With new 0.70.1 version it does not crash any more:-)

Sign In or Register to comment.