I have noticed that my OH1 needs a reboot every 70-ish days…
Looking into this issue revealed that Java has 40,000 files open…
sudo lsof | cut -d " " -f 1 | uniq -c | sort -nr
40041 java
415 apache2
...
Looking up the Java PID and listing the files with:
sudo ls -l /proc/9466/fd
shows tons of openHAB files.
How can I fix his?
Can I fix it? As in, is it an openAB problem?
What other info is needed to look into this?
Reboot the device. That will fix the problem. It might come back again, though.
As to whether it’s an OH problem, is harder to say.
You should make a list of the bundles that were running during the time period in question. It’s possible one (or more) of them is leaking.
Yes, reboot will fix it, but is a band aid rather than a solution. 
And it does come back…
I am interested in a fix; e.g. a proper way to get to the bottom of this, so that I can log an issue with whomever I need to…
Any hints appreciated…
Today’s stats look like this:
sudo lsof | awk '{ print $2 " " $1; }' | sort -rn | uniq -c | sort -rn | head -20
41269 9466 java
380 669 llvmpipe-
...
Again, we need to know which bundles/bindings were running.
Also, it would be interesting to compare the numbers above with these scenarios:
Hmmm…
This is from a fresh install of openHABian… not config done to it, nothing installed (other than lsof)
I selected EXPERT mode, hence, no bindings are installed as I understand it.
[12:35:40] root@openHABianPi:/home/openhabian# sudo lsof | awk '{ print $2 " " $1; }' | sort -rn | uniq -c | sort -rn | head -20
36618 31302 java
193 7012 smbd
188 29706 smbd
182 29712 smbd
156 30780 node
104 30780 V8
55 31470 sshd
51 31461 sshd
45 1 systemd
39 132 systemd-j
38 420 dbus-daem
37 417 avahi-dae
36 5486 systemd-u
34 477 wpa_suppl
33 844 sshd
33 539 bluetooth
33 10039 rpcbind
32 31954 su
31 31950 sudo
31 1058 sudo
[12:35:50] root@openHABianPi:/home/openhabian# uptime
12:38:30 up 7 days, 4:18, 1 user, load average: 0.11, 0.30, 0.18