APC_MODBUS _apc_modbus_read_registers Timeouts #2609
Comments
|
Build command for libmodbus and NUT build_libmodbus.sh #!/bin/bash
if [ "$EUID" -ne 0 ]; then
echo -e "ERROR: Please run as root\n"
exit
fi
# Install dependencies
apt-get install -y \
autoconf libusb-1.0-0-dev libusb-dev git libtool gawk
# Build RTU USB Library
git clone -b rtu_usb https://github.com/networkupstools/libmodbus
cd libmodbus
./autogen.sh
export DEB_HOST_MULTIARCH=aarch64-linux-gnu
./configure \
--prefix=/usr \
--sysconfdir=/etc/nut \
--includedir=/usr/include \
--mandir=/usr/share/man \
--libdir=/usr/lib/${DEB_HOST_MULTIARCH} \
--libexecdir=/usr/libexec \
--with-libusb \
--enable-static \
--enable-shared \
--with-pkgconfig-dir=/usr/lib/${DEB_HOST_MULTIARCH}/pkgconfig
make -j4 && make installbuild_nut.sh #!/bin/bash
if [ "$EUID" -ne 0 ]; then
echo -e "ERROR: Please run as root\n"
exit
fi
# Install dependencies
apt-get install -y \
libssl-dev libnss3-dev libgd-dev augeas-tools \
libaugeas-dev augeas-lenses libcppunit-dev \
ccache time perl curl make autoconf automake \
libltdl-dev libtool cppcheck pkg-config gcc g++ \
clang valgrind libsystemd-dev
# Build NUT
git clone https://github.com/networkupstools/nut
cd nut
./autogen.sh
export DEB_HOST_MULTIARCH="aarch64-linux-gnu"
./configure \
--prefix=/usr \
--with-doc=no \
--with-drivers="apc_modbus,usbhid-ups" \
--with-usb \
--with-modbus \
--sysconfdir=/etc/nut \
--includedir=/usr/include \
--mandir=/usr/share/man \
--libdir=/usr/lib/${DEB_HOST_MULTIARCH} \
--libexecdir=/usr/libexec \
--with-ssl \
--with-nss \
--with-cgi \
--with-dev \
--with-statepath=/run/nut \
--with-altpidpath=/run/nut \
--with-drvpath=/lib/nut \
--with-cgipath=/usr/lib/cgi-bin/nut \
--with-htmlpath=/usr/share/nut/www \
--with-pidpath=/run/nut \
--datadir=/usr/share/nut \
--with-pkgconfig-dir=/usr/lib/${DEB_HOST_MULTIARCH}/pkgconfig \
--with-user=nut \
--with-group=nut \
--with-udev-dir=/lib/udev \
--with-systemdsystemunitdir=/lib/systemd/system \
--with-systemdshutdowndir=/lib/systemd/system-shutdown \
--with-systemdtmpfilesdir=/usr/lib/tmpfiles.d
make -j4
echo ""
read -p "Do you wish to continue? [Y/N] " yesno
case $yesno in
[Yy]* )
echo "Installing..."
;;
* )
echo "Exiting..."
exit
;;
esac
make install |
|
CC @EchterAgo : does this ring any bells, please? |
|
In an effort to rule out the RPi being a factor, I recreated the build on an x86 system with Ubuntu 24.04.1 which results in: |
|
Thanks for the diligence about multi-platform check. RPi's USB chips were also suspects a few times. But the segfault was reported in several issues recently - so seeing it misbehave differently per platform is at least "most intriguing!" :) CC @EchterAgo again I might at best look at code in a couple of weeks or more, and have no similar APC device to check against. |
jimklimov
added
bug
APC
USB
modbus
impacts-release-2.8.2
|
@jimklimov Thanks for the reply. I will keep a look out for updates, let me know if there is anything you'd like me to try. |
|
To add another device: I'm seeing a similar behavior using a Smart-UPS 750 running FW 9.3 on NUT 2.8.2 with libmodbus c02f62b58aa4ce7538e087ef46266d2a048b0bd6. |
|
UPDATE: Seems like a practical duplicate of #2289 |
|
To summarize current findings in one place, we see several similar patterns ending up in driver crash (typically with
** Notably, Separately from this (maybe... seems similar to first sample above), in the later after-party discussion of original PR #2063 there were at least occasional (fixed by cable re-plug) faults with: And also an The I suppose a run with There are other FD macros nearby, all defined in terms of access to https://stackoverflow.com/questions/22827510/how-to-avoid-bad-fd-set-buffer-overflow-crash suggests there are problems with 1000+ socket/FD numbers, and also reminds to check if there are FD leaks (loops with a FD not closed so another number gets opened and issued for some next operation). Another question is what CC @EchterAgo |
|
I suppose That's about all I can say for now, lacking a device to test against, time (gotta run in fact), etc. But I really hope the data-dig above can get someone started! |
|
@jimklimov I can provide you with system. If you want, Please send me a mail. |
|
@ZerglingGo : thanks I will, but first... @ all : were all those attempts done with the proper NUT setup, drivers running unprivileged by default? That "Resource temporarily unavailable" clue above bothers me since I saw it yesterday. Can you please try running the driver as The hypothesis I'm trying to check here is if permissions for In the debug dumps posted earlier I see |
|
I guess that should be |
|
Sorry, yes. Thought about |
|
https://gist.github.com/ZerglingGo/1cc0d27ae30e26d4000d674b0f7da6ec#file-strace_apc_modbus-log-2 |
On an APC SMT1500C device using the rtu_usb version of libmodbus and a USB cable, reads fail with a timeout.
Since support for USB based APC Modbus is not mainlined, I am building the required RTU_USB version of libmodbus and compiling the main branch of NUT with usb and modbus support per the https://github.com/networkupstools/nut/wiki/APC-UPS-with-Modbus-protocol wiki page. The build appears to be successful but I am not able to at get the apc_modbus driver to communicate with the UPS.
I have attached build logs and included the apc_modbus output below. I have tested the connection using APC's PowerChute program (on a different system) and PowerChute is able to collect all of the ModBus only fields (like load and temperatures, etc). So I am fairly confident it is not a physical connection issue. Also when using the standard Debian bookworm build of NUT I am able to reliably communicate using the usbhid-ups driver but as it only provides a few fields, the Modbus driver is prefered.
Any pointers would be appreciated but I am not sure where to go from here.
NUT configuration indicating the new RTU USB features are included in the NUT build
Using apc_modbus to probe the unit this is the output:
Host Details
ups.conf
upsd.conf
nut.conf
The text was updated successfully, but these errors were encountered: