"error in read()" error messages about mbusd HOT 7 CLOSED

Many thanks for the quick turn around.

Will try to test a fix next week.

The actual system in question is several thousand miles away...
thankfully ssh makes that not too much of a problem!

from mbusd.

Apologies for the late response - the well the device was on was shut for a several weeks. System turned off.

We've been using the fix on some other sites and have not seen the original error messages there.

So i would vote to close the issue. Many thanks for the fast response and fix!

from mbusd.

One more datapoint...

We have also seen those errors with errno set to EFAULT eg.

mbusd[25373]: 30 Oct 2021 13:30:40 tty: error in read() (Bad address)
mbusd[25373]: 30 Oct 2021 17:38:18 tty: error in read() (Bad address)

from mbusd.

@AlexDavies501 thanks for the report! Could you provide some additional info about the software and hardware environment you're running? Kernel version, platform, etc.

from mbusd.

Hi.

Edited uname -a output (removed the computer name) below:

Linux datahub 4.1.15-g33f8ed0f5cb3 #4 SMP PREEMPT Tue May 1 13:54:29 MST 2018 armv7l GNU/Linux

The hardware is a Technologics 7970 embedded computer. Has two serial ports. We run mbusd against one of these.

/usr/local/bin/mbusd -d -v2 -L - -c /usr/local/etc/mbusd/mbusd-ttyS1.conf -p /dev/ttyS1

The config file now looks like:

# Serial port device name
device = /dev/ttyS1

# Serial port speed
speed = 9600

# Serial port mode
mode = 8n1

# RS-485 data direction control type (addc, rts, sysfs_0, sysfs_1)
trx_control = addc

# Sysfs file to use to control data direction
# trx_sysfile =

############# TCP port settings #############

# TCP server address to bind
address = 0.0.0.0

# TCP server port number
port = 12502

# Maximum number of simultaneous TCP connections
maxconn = 32

# Connection timeout value in seconds
timeout = 60

######### Request/response settings #########

# Maximum number of request retries
retries = 0

# Pause between requests in milliseconds
pause = 100

# Response wait time in milliseconds
wait = 500

With the above config i see a dozen tty: error in read() (Resource temporarily unavailable) errors for today, and 3 or 4 for each of the previous 4 days.

Originally, when we saw a lot more of these errors (several hundred per day), the config had retries=3 and wait=1000.

There are 3 regular callers that create Modbus master connections to the TCP side of mbusd. These tend to do a poll of a handful of registers every 5 seconds. Also some occasional scripts manually called from the cmdline that make connections too.

There are two slave devices on the 485 network: a gateway device and a pressure transmitter.

from mbusd.

Hi, thanks for the info!

I've pushed some fixes regarding handling EAGAIN read() result. Сould you test mbusd compiled from the latest sources?

from mbusd.

Glad to hear the problem is sorted out! So I'm closing this issue, please feel free to reopen if needed.

from mbusd.