WRT3200acm Version 1
Yesterday I updated to
Lede SNAPSHOT r7210-14cb05909a
Everything works as before, but Privoxy stopped working. Was fine on prior LEDE on this router
Have been using Privoxy with OpenWrt / LEDE on WRT1900AC (ver 1) never problems
Deleted / Reinstalled Privoxy: No Dice same settings as before 192.168.4.1:8118 no luck
P.S. Has anyone installed Privoxy compiled for Compression? I would love to give it a shot, but I'm not
up on compiling (last time was MS BASIC 7.1 for DOS & OS/2 in 1990s)
Thank's David for an EXCELLENT Build!
Mariano - Please accept my apologies for not being more specific in my response to the post by directnupe. The issue I experienced was in fact with the Material theme. I personally do not like the Dark Matter theme due to its aesthetics.
After I upgraded to the newer build using the .bin file, do not keep settings, the buttons on the startup page were stacked on top of each other and did not seem to be rendering properly. Cleared my caches, rebooted and it was still there. At first I ignored it, but when I discovered the issue with not being able to change the LAN interface, I reverted back to my previous version and all was fine.
Not sure if this issue is specific to mamba devices but it would seem so as directnupe and I both use the same router.
~Tala~
Mariano - Thanks for your response. Aside from the additional config tweaks, that was actually the main reason, not having automated updates. In hindsight, your comment on using a one-year old blacklist made a lot of sense to me so I decided follow your lead and switched over to using dnscrypt-proxy for adblocking. I installed the public blacklist from https://download.dnscrypt.info/blacklists/domains/ and it is working flawlessly. Not sure if one can specify multiple blacklist files in the .toml configuration file but I would also like to add the domains-blacklist-local-additions.
Thanks again for all the help.
~Tala~
Still struggling with ping/latency but just wanted to chime in with a quick update on the above: it seems the Amazon equipment (Dot, Echo and FireTV at least) have a lot of problems with DHCP. I've found loads of threads with respect to these devices on Microtik and other open source router platforms. It seems Amazon's DHCP client implementation is buggy as hell and the recomendation in these instances is to give the devices static IPs.
I thought you had a separate device to handle DHCP requests, so that the IP allocation mechansim would hvae been the same with both the apple airport extreeme and the wrt32x as your gateway.
Yeah, I've been experimenting with using the built-in DHCP server and my LAN server to see if that was causing issues with the LAN connectivity. When running the DHCP server on openwrt I see loads of cycles of DISCOVER and OFFER but never any REQUEST or ACK from the Echo and Dot devices. Very odd behaviour that seems to be sorted by disabling the openwrt dhcp instance on the LAN interface and using my server instead.
Thanks for this build. However, I will never use UPnP, so I figured I'd remove the relevant packages. I think I'd only remove these two to remove UPnP and it's associated LuCi GUI:
luci-app-upnp
miniupnpd
Do I have that correct or are there more to remove?
I would just leave it there, it won't hurt if you dont use it...
Im using UPNP for my Game consoles (and only for them!!), works just fine and so far i do get the open NAT status in every game !
Okay, this will hopefully be the last of my posts re. wrt32x. It's working!
I finally had the time today change all of my LAN to 192.168.1.0/24, including my server and the few static IP devices I have configured. I put the latest davidc502 build back on the wrt32x, booted it up, turned off dhcpd on my server and everything works. EVERYTHING.
All devices are happily getting DHCP leases first time, pings and jitter are back to normal, wifi stations are disconnecting after 300s inactivity rather than 5.
I'm stumped, I really am. There's obviously something deep inside that really doesn't like the change from 10.0.0.0/24 to 192.168.1.0/24, something's not getting updated properly along with the NIC change. I dunno what, but it's working now and I'm too tired to debug anymore to find out why. That's for another day...
Congrats and that's good news. Glad to know there is little to fix on davidc502's builds. I've always run my network on 192.168.1.0/24 so haven't seen this issue. The high latency issue really concerned me because I do game where the lowest latency/best network performance is paramount.
I managed to get through to Virgin Media's "go to person" to downgrade/upgrade from a Super Hub 3 to a Super Hub 2ac to avoid any latency issues. Pinging the modem would show so much erratic behavior that it was well worth doing. His issue was concerning me too until finding this out.
Build is good here, no latency issues. Using 192.168.0.1 as usual.
Not sure if anyone else can reproduce this with r7210 on the WRT32X, by initiating a ping -t google.co.uk command in Windows 10 causes a spike in CPU usage 0.05 to 0.70. It goes back to normal eventually.
Not on Win10 here (macOS 10.13.5) but pinging google.co.uk does nothing to my WRT32X. Worth pointing out I've disabled dnscryptproxy, though, just running straight dnsmasq using Google for upstream name resolution.
Recently, I switched from r5621 to r7210 on my wrt3200acm.
This entailed my avahi installation being upgraded from v0.6.32 to v0.7-1.
In v0.7-1, it seems that the daemon wants to use '/run/' instead of the normal '/var/run/' as runtime directory for the socket etc.
Which runtime directory to use is defined build time and can't be modified by any option for the daemon.
A change was introduced in v0.7.
A fix seems to be on the way.
While davidc502-builds still use avahi v0.7, a workaround could be to create '/run' instead of '/var/run/' from the startup script (/etc/init.d/avahi-daemon). But I'm worried about putting the resulting daemon socket on flash since the '/run' folder will end up on the overlay, i.e. flash memory.
Question:
With the lifetime of a flash memory in mind.
Is is safe to have a socket created in the userdefined root-folder '/run/...' instead of the standard '/var/run/...'?
