Thanks, puzzling... as another datapoint, would NUT's `make check-NIT` behave reasonably? It is a set of (hopefully portable, not just bash) scripts to set up and run upsd vs upsc locally. Would at least that work?..
Also, would the older NUT 2.7.4 release fare better? It is pre-refactoring that happened in the past couple of years to reduce warnings (including data type mismatches) - maybe some int size behind file descriptors popped out wrong here? Jim On Mon, Dec 19, 2022, 17:41 vom513 via Nut-upsuser < [email protected]> wrote: > > > On Dec 19, 2022, at 9:25 AM, Jim Klimov <[email protected]> wrote: > > > > Maybe also try strace, truss and the likes to see what syscalls it > makes? e.g. falls on read/select?.. > > > > Yep, it’s not even trying a connect(). Here is upsmon vs. curl: > > [root@ss20 ~]# LD_LIBRARY_PATH=/usr/local/ups/lib truss -f > /usr/local/ups/sbin/upsmon 2>&1 | grep -i connect > Broadcast Message from nobody (???) on ss20 Mon Dec 19 11:19:26... > Communications with UPS [email protected] lost > Broadcast Message from nobody (???) on ss20 Mon Dec 19 11:19:31... > UPS [email protected] is unavailable > ^C > > [root@ss20 ~]# truss curl -I -v http://burn.net 2>&1 | grep -i connect > > About to connect() to burn.net port 80 (#0) > connect(4, 0xEFFFEC98, 32, 1) Err#150 EINPROGRESS > connected > Connected to burn.net (2001:19f0:5c01:19a3:45:76:18:80) port 80 (#0) > Closing connection #0 > > > > _______________________________________________ > Nut-upsuser mailing list > [email protected] > https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser >
_______________________________________________ Nut-upsuser mailing list [email protected] https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/nut-upsuser
