did you find a fix, this just stopped me from installing technetium as a docker app, sigh
whats odd is dnsmasq is listening on a different interface to the docker interface, so this really shouldn’t fail…
Turns out you are able to specify an interface for containers if you convert them into a yaml file/install a custom instance with a yaml. (Didn’t realize this was an option in TrueNAS because it was hidden lol)
It looks like this in the yaml
ports:
- "interface-ip:port:port"
I specified my default interface for Pihole.
Then I created a virtual interface called virt0 and edited the dnsmasq config for Incus.
/etc/dnsmasq.d/incus
# Tell any system-wide dnsmasq instance to make sure to bind to interfaces
# instead of listening on 0.0.0.0
# WARNING: changes to this file will be lost if incus is removed.
bind-interfaces
interface=virt0
except-interface=incusbr0
Hi All,
Not sure if this is a new issue, but I am using zabbix to poll Truenas and I can see a lot syslog messages like this:
<30>1 2025-03-01T19:09:15+01:00 TrueNAS snmpd 63506 - - systemstats_linux: unexpected header length in /proc/net/snmp. 237 != 224
Probably self-stupidity on my part, and I read the migration process of VMs (I don’t care I can recreate them if they are lost) but after upgrading to this version from 24 there are no VM menu options available.
Incus is apparently installed, the upgrade went prefect, etc. No biggie…just can’t make VMs with the interface.
It’s under instances now.
lol, just found it literally 10 seconds ago and was going to reply I am in fact an idiot.
It would be good to start a new thread on this in general.
Can you verify that zabbix had no problems with 24.10?
Is there another SNMP manager with similar issues?
After a few days of testing, no more issues that affect usage were found except for those that are known to be fixed in RC1.
In my test, TN 25.04 has surpassed TNC 13.0 in performance. I am looking forward to the release of RC1.
3 Likes
Out of curiosity, what kind of perf test did you do to compare the two?
Core 13.0? That’s using a version of OpenZFS two major releases behind.
The server uses ryzen 9700x, which is undoubtedly one of the best single-core performance CPU currently available. The performance test results of 25.04 are equivalent to or even higher than 13.0.
Some examples:
Use fio on host test how much read and write performance “operating only data in arc” can provide (–size=10M)
Use fio on Windows to test smb peak performance
(–bs=1M) reaches 6GiB/s read and 3.8GiB/s write
(–bs=4K) reaches 100k iops read and write
This is the test result of a single windows smb client, and multiple clients should be able to get higher.
1 Like
ZFS performance itself will not have much impact.
current performance upper limit is on each file/block transfer protocol.
The 9700X isn’t even in the top 10 in single core performance. What are you basing that on? It’s not even the best AMD desktop CPU for single core performance. The 9700X single core performance is similar to a 3 year old mid range intel CPU.
Yes, lets make this thread a flamwar on AMD/Intel
2 Likes
FYI…
The 25.04-RC.1 version is going through QA and looking good for next week.
Thanks to all the BETA testers who have provided feedback and bug reports.
8 Likes
SMB shares no longer accessible for existing users. “NT_STATUS_NO_SUCH_USER”
Newly created users can access shares.
There are two issues in beta related to SMB users:
- if for some reason you have users who don’t have an smbhash in the DB (for instance after restore without a secret seed), the SMB user synchronization raises an error and does not complete.
- if for some reason a user overrode our default of making usernames all lowercase (eww) then the user isn’t added properly to the SMB passdb.
Both are fixed for RC.
3 Likes
Uppercase characters would explain it!
Have one user way down the list that works and their name is all lowercase.
And this is why we have a non-essential test machine!
Wasn’t a flame war, (not sure what a flamwar is) but a claim was made, and I was wondering where it came from. How about not posting unless you have something to add. 
Agreed, but along the same lines, maybe you could then explain what 3 year mid-range intel CPU that you think Ryzen 9700x would be equiv to, single core?