I have a 4-5 year old fully "updated". Samsung Tizen based TV. I have also factory reset this tv and started over with the configuration.
that's means removing any additional or any apps I don't want. It has Apple TV, amazon, and YouTube app, Samsung TV.
its on its own separate Vlan/ dedicated port on my DEC 740 Appliance.
what are your thought that would cause memory to increase by X2 by just turning on a "smart" device? it pretty much stays that high even up to 30 minutes after the tv is turned off?
Lots of firewall states? Do you see a big increase in Reporting->Health->System (states) for that time period?
will check the next time I turn on the tv.
I am probably going to get a different product to stream. my gut feeling is tizen is no longer supported / EOL
was hoping for someone with more knowledge to find where its coming from thanks opnenthu
Quote from: DEC740airp414user on May 03, 2026, 12:07:27 PMwas hoping for someone with more knowledge...
Hey no problem. :) Take care.
My Hisense telly likes to call its friends around the world. No updates, adfest on TV channels. Are you running Unbound and is it simply caching all the websites your telly also likes to invite?
Quote from: DEC740airp414user on May 03, 2026, 10:50:25 AMwhat are your thought that would cause memory to increase by X2 by just turning on a "smart" device? it pretty much stays that high even up to 30 minutes after the tv is turned off?
Considering, the TV only reaches to internet. The only thing how it can impact memory is as mentioned above by opening a lot of sessions.
Each session takes a slice of memory. If you run ZA or IPS the memory impact can be multiplied.
Quoteeach state table entry requires about 1 kB (kilobytes) of RAM
Considering the above, the TV would have to open a lot of sessions. But SmartTVs are basically just IoT devices with the perk they like to flood and spam which is as well multiplied by various applications you run on the TV.
Regards,
S.
Now I'm really puzzled
I disconnected my tizen tv from the internet... Bought a brand new Apple TV 4K and hardwired it to the same port in my dec appliance
It's doing the same thing
Doing the health - system - states pfstates goes from 0 to 450.
In utilizing unbound with
Adguard
Oisd large
Hagazi pro
And the pro version of Qfeeds
😳
0 to 450 is a significant increase but 450 is "nothing". Why would you worry about 450 sessions? Thousands of them is what I would consider normal even in a home network with one or two dozens of devices.
Quote from: Patrick M. Hausen on May 03, 2026, 10:57:30 PM0 to 450 is a significant increase but 450 is "nothing". Why would you worry about 450 sessions? Thousands of them is what I would consider normal even in a home network with one or two dozens of devices.
I was just following the suggestion from a post above. to try to find why memory doubles in usage after a device turns on
I've had this same TV and DEC 740 router for months now. the memory change didn't happen it seems until 26.4 I am guessing. I am open to more suggestions on why this is happening.
Login to your OPNsense via SSH.
Invoke "top".
Type 'o' followed by 'res' and ENTER.
Power on the device.
Check which process is using that memory.
seems to be python and unbound. unbound always takes up over a 1gb of ram for me
was out of upload size space
When it comes to Samsung TVs I am not surprised : https://discourse.pi-hole.net/t/this-is-what-your-dashboard-looks-like-with-a-samsung-tv/3165 ;)
And in general anything Samsung has a lot of pre-installed crap on it so you should uninstall or disable as much as stuff as possible anyway ;)
Another thing you could look at is how much EULA stuff you REALLY NEED in order to have a functioning TV : Usually you don't need all of them and the ones you can avoid are also usually the ones that deliver you the most advertising crap !!
@DEC740airp414user use "ps awwwux | grep python3" to see what these python processes are doing.
root@OPNsense:~ # ps awwwux | grep python3
root 390 0.0 0.0 32452 4 - IWs - 0:00.00 /usr/local/bin/python3 /usr/local/opnsense/service/configd.py (python3.13)
root 392 0.0 0.7 177528 29712 - I Fri06 0:19.12 /usr/local/bin/python3 /usr/local/opnsense/service/configd.py console (python3.13)
root 9430 0.0 0.3 28620 14208 - S 12:23 0:00.63 /usr/local/bin/python3 /usr/local/sbin/configctl -e -t 0.5 system event config_changed (python3.13)
root 50783 0.0 3.2 3502492 132300 - S 12:23 0:26.78 /usr/local/bin/python3 /usr/local/opnsense/scripts/unbound/logger.py (python3.13)
root 66912 0.0 0.2 28632 7640 - S Fri06 0:19.88 /usr/local/bin/python3 /usr/local/opnsense/scripts/syslog/lockout_handler (python3.13)
root 58074 0.0 0.0 13744 1964 0 S+ 14:15 0:00.00 grep python3
root 50783 0.0 3.2 3502492 132300 - S 12:23 0:26.78 /usr/local/bin/python3 /usr/local/opnsense/scripts/unbound/logger.py (python3.13)
Do you have query logging enabled in Unbound? Probably better don't. Unbound is a robust and secure recursive name server but not a scalable filtering solution. If you use large block lists and want to log queries, consider using AdGuard Home for that and Unbound as AGH's upstream.
HTH,
Patrick
I wondered that was well and double checked I didn't. Unless it's somewhere else than the advanced tab?
Then check the log for WTH it is that blows up the logger script to this size. Maybe the TV is generating a crap ton of bogus queries? No idea, seriously. My statement about scalability still holds, IMHO. AGH or PiHole perform much better than Unbound because Unbound wasn't designed for really large dynamic filter lists.
I turned off all but Qfeeds blocklist for unbound and turned this off under reporting - settings.
I'm forwarding to controld/ nextdns anyways
Thanks for the replies Patrick