search

osrf / srcsim

Space Robotics Challenge
Other
9 stars 3 forks source link

Stability issues #161

Open osrf-migration opened 7 years ago

osrf-migration commented 7 years ago

Original report (archived issue) by Víctor López (Bitbucket: Victor Lopez).

If we start the simulation, and leave it be without us sending any command or starting any other software, the robotics will eventually fall, usually on the first 30min of simulated time.

Is anyone seeing this too? How are we supposed to run tasks for 3h of simulated time?

Will be it help if we provide rosbags and gazebo logs of it?

osrf-migration commented 7 years ago

Original comment by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

Yes, rosbags and Gazebo logs should help. Thanks!

osrf-migration commented 7 years ago

Original comment by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

osrf-migration commented 7 years ago

Original comment by Jeremy White (Bitbucket: knitfoo).

@vlopez87: is this a duplicate of bug #134?

Beyond #134, which I think I've mitigated, I am still seeing the very occasional 'freak collapse'. I haven't worked out a pattern, or a real rhyme or reason to it as of yet. It seems to be of two main forms. The first is a collapse at start, which I think is somewhat described in #98 (although I get the sense that the people in #98 see it consistently, whereas I see it only sporadically).

The second form I see is a bit more rare, and is similar to #134, but I've seen it happen after I upped my file handle limit, and the failure mode is different than that of #134. That is, the robot just collapses, usually in the middle of a routine walk.

I've scrutinized all the logs I had gathered, and haven't seen anything. I've collected one example and will attach it in case someone else notices something I've missed.

osrf-migration commented 7 years ago

Original comment by Víctor López (Bitbucket: Victor Lopez).

When it falls we don't necessarily get the Too many open files error. Sometimes it's an "Address already in use", even though we just rebooted the computer.

I'll be recording gazebo logs and rosbags and will try to post them when we see the failures again.

osrf-migration commented 7 years ago

Original comment by Vinayak Jagtap (Bitbucket: vinayak_jagtap).

We are seeing stability issues while walking when the entire world is loaded and real time factor is below 0.1. The planned footsteps are executed just fine when we load only task1 or task2. Haven't yet tested with the pull request #67