Hi!
Do you know why system log shows this very often?
daemon.warn dnsmasq[4071]: failed to create listening socket for 127.0.0.1: Address in use
I don't know if this is the reason that is causing DNS queries sometimes takes longer.
BTW, I'm using dnscrypt-proxy2.
Thank you!
Look at the output of:
cat /etc/config/dnscrypt-proxy.toml | grep listen_addresses
It should return:
## listen_addresses = ['127.0.0.1:53', '[::1]:53']
listen_addresses = ['127.0.0.53:53']
# listen_addresses = ['127.0.0.1:3000']
Also cat /etc/config/dhcp | grep 'list server' which should return:
list server '127.0.0.53'
I suspect you will see one or both returns will be 127.0.0.1
Any plans on new build? I thought people said a bug is causing builds to not work anymore but https://downloads.openwrt.org/snapshots/targets/mvebu/cortexa9/ shows a build from build happened successfully.
http://builds.openwrt.org/master/images/builders/mvebu%2Fcortexa9 says its building again since 3 days.
At this point there are no plans for any more builds.
And why is that the case? This memory leak issue of dnscrypt-proxy2 seems to be fixed. This is reason enough to update a build. Do you have stopped the project? You did a build twice a month, and over 2 months of no build, it is obviously a good idea to make one if it works again to build. Of all the package updates alone.
It has been explained multiple times in this thread.
I for one disagree. Might I humbly suggest that the time has come for you to investigate building your own to your specific wants?
There is a memory leak, it has apparently been fixed, and I have previously confirmed your suspicions here, but to demand a new build to resolve such a minor issue seems a bit much to my mind when a simple cron to reboot in the wee hours of the morning makes your problem go away.
What are you even talking about, no it was never explained, the explanation was that builds are broken, but they are not anymore since 3 days, so there is no meaning in not building again. Memory leak which crashes the router is not a minor issue.
Oh, I was hoping to see a new build in the near future... 
Anyway I hope that you will change your mind and keep your great work up. I have much respect for the work that you have done so far!
Best Regards.
My (tic) point was that ya canât build from a daily that isnât being produced.
Agreed, but come on. It took what? 61 days to finally die!
Please remember Snapshot builds are not stable. That David has managed to produce successive, very stable builds for these platforms is a pretty amazing feat in my view.
Iâve had my own issues with a broken softflowd/proc.d process in this build, but with some good help here, Iâve come up with a functional workaround - Problem Solved.
Iâve given you a real world solution to the problem that wonât cause any problems for anyone else (Itâs even a C/P for that matter), as well as another much more involved solution.
Those are pretty much your only options atm.
Edit:
I just thought of one other alternative for you. Download the âfixedâ version, extract and compress the binary per the Read Me, replace the current binary with the compressed one, restart the service. The binary compresses from ~8MB -> ~3.5MB and runs just fine.
There is a cost to this of course in that you will not gain anything in /overlay and it needs at least 4MB of free /overlay to do it.
Much appreciated..
Everything else aside, I've been compiling OpenWrt/LEDE I think for well over 5 years, and over time interests change. There is a new project coming out not related to routers, and I think I want to make my own branch and start developing the the way I would like. With covid-19 still going on who knows when the project will be available (new product), but will start ramping up when the time comes.
Edit Who knows what will happen, but I'm just going to wait for some of the current problems I'm reading about and we know about to pass, and I'll compile more builds and make them available. As if I don't already have much of a presence, I'll probably have even less.
David, it might be worth adding a paragraph to your main site (https://dc502wrt.org/) explaining that the 5/24 is the last build for an undetermined length of time because of DSA soft-bricking everyone's routers and that daily builds aren't being produced (or whatever the actual reason is, I may not have stated it correctly). Perhaps that will help mitigate the question occurring in this thread.
Thanks... I'll do that.
Updated.
I've been following this thread and your builds for quite sometimes and want to thank you for all your work, you just made my router more happy:). I understand the problems DSA is causing and settled myself on using your last build (r13342) which is very stable until something come up and all the bugs are mostly fixed with DSA. Hopefully, I'll see more of your custom builds in the future. Thank again and god bless you.
Can't say I've noticed any bugs / issues with DSA post the initial fire-fights in the PR which introduced this functionality. I have had few missteps as I have tried various configuration / setup methods over the last couple of months of use, but I don't consider user error a bug.
Hi!
Thank you for your response.
This is what is shown with your recommendation:
## listen_addresses = ['127.0.0.1:53', '[::1]:53']
listen_addresses = ['127.0.0.1:53', '[::FFFF:127.0.0.53]:53']
## listen_addresses = ['127.0.0.53:53']
# listen_addresses = ['127.0.0.1:3000']
Is there something wrong?
Thank you!
Are you using IP6. If so change the
listen_addresses = ['127.0.0.1:53', '[::FFFF:127.0.0.53]:53']
to
listen_addresses = ['127.0.0.53:53', '[::FFFF:127.0.0.53]:53']
In your dnscrypt-proxy.toml file. If youâre just using IP4, comment the above line out, un-comment the next line and make sure the IP is 127.0.0.53:53.
You didnât post what is in your dhcp file, but just make sure that the
list server '127.0.0.53'
entry in dhcp file under dnsmasq section matches above.
After doing that, System log shows this:
daemon.err dnscrypt-proxy[5628]: [2020-07-31 03:45:18] [FATAL] listen udp 127.0.0.53:53: bind: address already in use
And DNS fails completely.
If set to 127.0.0.53 on both addresses, fails, if set to 127.0.0.1 on both addresses fails too. So the only way it works for me is how i showed you.
Leaving as it was, DNS works fine on IPv4 and IPv6, but a bunch of messages are shown all the time:
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:01 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:03 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:07 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:07 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:07 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
Thu Jul 30 23:01:07 2020 daemon.warn dnsmasq[4380]: failed to create listening socket for 127.0.0.1: Address in use
What i'm missing?
Thank you!!!!
cat /etc/config/dhcp
and post just the config dnsmasq section.
By default dnsmasq listens on 127.0.0.1:53 but it complains under your original setup that another process was already using it.
This is:
config dnsmasq
option domainneeded '1'
option localise_queries '1'
option rebind_protection '1'
option rebind_localhost '1'
option local '/lan/'
option domain 'lan'
option expandhosts '1'
option authoritative '1'
option readethers '1'
option leasefile '/tmp/dhcp.leases'
option resolvfile '/tmp/resolv.conf.d/resolv.conf.auto'
option localservice '1'
option confdir '/tmp/dnsmasq.d'
option noresolv '1'
option localuse '1'
option boguspriv '1'
list server '127.0.0.53'
Hope this helps.
Thanks!
I see nothing wrong there. Did you restart the dnscrypt-proxy and dnsmasq services after the changes?