Closed zhouchengming1 closed 6 years ago
Hi,
did you synchronize time between machines A and B? http://wiki.ros.org/ROS/NetworkSetup#Timing_issues.2C_TF_complaining_about_extrapolation_into_the_future.3F
@atiderko No, I will try this. But I don't know why /tf topic synced to B machine could cause a problem like this? Does it get some time information from B when register?
Does the multimaster sync node will modify the timestamp of local ros master ?
master_sync
does not modify any timestamps or other data or behaviour of ROS communication.
If you synchronize /tf (and also other topics) the sender stay the same and publish tf messages from different hosts. The receiver of /tf get messages with different timestamps.
I do not know if it will cause the BufferCore :: getLatestCommonTime error. It is just a guess!
If you use node_manager
you can temporary synchronize the time and check if the error occurs with synchronized time.
It seems that two rosnodes A and B are synced to another machine, A publish to /tf topic and B subscribe to /tf topic , so another data connection created on the remote machine? When I sync only needed topics, instead of the whole nodes, the problem is gone.
Hello, all:
I use multimaster to sync some topics and nodes for two machines, but I found if I sync /tf topic from A to B machine , then the program on A will get wrong time from BufferCore::getLatestCommonTime. I don't know how it happens, anyone have idea? Thanks.