First use of serial device throws "device already locked" exception about ecl_core

[AlexReimann]

Nope, what i meant was that some nodes of minimal go down and the robot is
not driveable
On Jan 2, 2015 10:58 PM, "Marcus Liebhardt" [email protected]
wrote:

After (re-)starting the machine, the first attempt of use a serial device
throws this exception:

ecl_core/ecl_devices/src/lib/serial_pos.cpp

Line 145 in 1337f3d

StandardException(LOC,OpenError,std::string("Device is already locked. Try 'lsof | grep ") + port + std::string("' to find other processes that currently have the port open (if the device is a symbolic link you may need to replace the device name with the device that it is pointing to).")));

On the second attempt (restarting the process (not the machine) using the
serial device) it works.

@AlexReimann https://github.com/alexreimann is this the error
(exception) you have been seeing, when you said "you have to start minimal
twice"?

—
Reply to this email directly or view it on GitHub
#41.

[stonier]

Let me know what errno. shows if you see it - we can look up the man page for fcnt to get the hint.

[AlexReimann]

It doesn't give an error code, it is just some ros nodes dying (during
the start of minimal).
On the second try everything works fine. Sometimes everything works on
the first try (2 or 3 times out of 50).

I actually never saw it happen on gocart 1.1.
There was no time for me to further investigate on wednesday because the
reliability test had priority.

Am 02.01.2015 um 23:05 schrieb Daniel Stonier:

Let me know what errno. shows if you see it - we can look up the man
page for |fcnt| to get the hint.

—
Reply to this email directly or view it on GitHub
#41 (comment).

[bit-pirate]

Just figured something out here.

1.0 was still outstarting the demo setup (minimal + dslam_navigation +
...). Need to run some tests with 1.0 to make sure this was the cause of
this issue. If it is the case, then the exception "device already locked"
would be correct -> 2x the same node tries to lock the USB/serial device.

On 2 January 2015 at 07:18, AlexReimann [email protected] wrote:

It doesn't give an error code, it is just some ros nodes dying (during
the start of minimal).
On the second try everything works fine. Sometimes everything works on
the first try (2 or 3 times out of 50).

I actually never saw it happen on gocart 1.1.
There was no time for me to further investigate on wednesday because the
reliability test had priority.

Am 02.01.2015 um 23:05 schrieb Daniel Stonier:

Let me know what errno. shows if you see it - we can look up the man
page for |fcnt| to get the hint.

—
Reply to this email directly or view it on GitHub
#41 (comment).

—
Reply to this email directly or view it on GitHub
#41 (comment).

Innovation Team Leader
Yujin Robot
Seoul, Republic of Korea
Twitter: @yujinrobotinno
Website: http://inno.yujinrobot.com
Email: [email protected]
Phone: [phone]

[stonier]

The posix error value will validate that.

[bit-pirate]

Haven't encounter this in a while now. Closing it. Re-open, if the issue reappears.