Re: [Emc-users] Machine slowed to a crawl
After rebooting, I restarted my job and it completed fine (it took 8 hours :( ) without using linuxcncrsh. Hmm, one possibility is there's a memory leak, and it ran the CNC system out of free memory. Was the disk busy (swapping)? But, I'm just guessing. I don't think so. The machine was very responsive. When running normally, axis uses most of the CPU so all other tasks run slowly. Are there some performance statistics I can check for if this happens again? I know the windows environment better - I know where to look for process memory leaks, resource leaks, etc. -- This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
[Emc-users] Machine slowed to a crawl
I originally posted this on IRC, but my network died so I couldn't discuss it. I was 4 hours into a 3D print using linuxcnc, and the machine crawled to an almost complete stop. My job was still running, but it executed 1 line of gcode, then pausing for 2 seconds. Axis wasn't showing the preview. The DRO read all 0. 'top' showed CPU at about 5% (it normally 80% or more because of axis). While the job was running, I was writing a monitoring application (on windows in c#) that would connect to linuxcncrsh and query the machine status every 5 seconds. I noticed the linuxcncrsh code reuses the nml interfacing code in halui.cc. Does this mean that linuxcncrsh uses the same nml channel as axis? Could this have been my problem? After rebooting, I restarted my job and it completed fine (it took 8 hours :( ) without using linuxcncrsh. -- This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
Re: [Emc-users] Machine slowed to a crawl
Frank Tkalcevic wrote: I originally posted this on IRC, but my network died so I couldn't discuss it. I was 4 hours into a 3D print using linuxcnc, and the machine crawled to an almost complete stop. My job was still running, but it executed 1 line of gcode, then pausing for 2 seconds. Axis wasn't showing the preview. The DRO read all 0. 'top' showed CPU at about 5% (it normally 80% or more because of axis). While the job was running, I was writing a monitoring application (on windows in c#) that would connect to linuxcncrsh and query the machine status every 5 seconds. I noticed the linuxcncrsh code reuses the nml interfacing code in halui.cc. Does this mean that linuxcncrsh uses the same nml channel as axis? Could this have been my problem? After rebooting, I restarted my job and it completed fine (it took 8 hours :( ) without using linuxcncrsh. Hmm, one possibility is there's a memory leak, and it ran the CNC system out of free memory. Was the disk busy (swapping)? But, I'm just guessing. Jon -- This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev ___ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users