when trying to access another computer in the same subnet by it's name (from within windows explorer), this is very unreliable:
- sometimes it works fine
- sometimes i need several retries
- sometimes i need to wait for several minute retrying again and again
- sometimes, it does not work at all
accessing the computer by it's IP address instead seems to works fine.
so this looks like for some reason, local machine names are not translated to IP addresses reliable. i have this issue since i use opnsense.
probably, i forgot to set a tick somewhere to let unbound dns translate host names in the local network.
unfortunately, i cannot find where to enable it.
I can second this.
While it worked flawlessly with 21.7.3, now with 21.7.5 sometimes the client gets the IP assigned, it also appears as such in the DHCPv4 leases list (together with its hostname), but it is not present inside /var/etc/dnsmasq-hosts. In that case other clients then cannot resolve the name.
for me, this did not start with 21.7.
i already had this issue with 20.7.8, but did not ask so far.
but i think it really got worse since my first installation of 20.7 about a year ago, maybe after doing some updates (but i'm not shure about this).
Wow, that's strange ... and you probably won't believe me this ... without doing anything - just after waiting 2 hours and trying again - the name is now resolvable ... and it is also inserted into /var/etc/dnsmasq-hosts where it wasn't 2 hours ago.
I have also noticed this, on my imac. I assumed it was an issue with the imac, flushing the imac dns did not help.
For me it becomes apparent when I add another VM, and usually have to resort to using the ip address, then after some time I can use the fqdn, or the short version.
This has been occurring for at least a year or more.