Posted: Thu Oct 10, 2019 16:17 Post subject: Netgear X10 r9000
Router/Version: Netgear X10 r9000
File: Firmware: DD-WRT v3.0-r41269 std (10/10/19)
Kernel: Linux 4.9.196 #390 SMP Thu Oct 10 01:45:02 CEST 2019 armv7l
Status: Operational (with issues)
Upgraded from: r41218
Issue #1: Privoxy @ ip 10.0.0.1:8118 returns error: "Maximum number of open connections reached." but config.privoxy.org does work.
Issue #2: SAMBA is not started automatically upon reboot. It is enabled but stopped. Need to 'apply' then 'save' 2 times to make it stick.
Issue #3: Syslog shows a few errors such as the following:
Oct 10 12:00:23 r9000 daemon.warn radvd: exiting, 1 sigterm(s) received
Oct 10 12:02:44 r9000 daemon.warn dnsmasq: warning: ignoring resolv-file flag because no-resolv is set
Oct 10 12:06:05 r9000 daemon.warn dnsmasq: nameserver 127.0.0.1 refused to do a recursive query
Oct 10 12:06:05 r9000 daemon.warn dnsmasq: possible DNS-rebind attack detected: cooper.logs.roku.com
RED Warnings: NONE so far in this build!
Memory usage appears to be much better in this build and will update with new information if needed.
Routers: TP-Link TL-wdr4900v2
Upgrade: via web
Firmware: v3.0-r41269 std (10/10/19)
Kernel: 3.18.140 #56708 Thu Oct 10 05:09:59 +04 2019 mips
Previous: 41243 private build
Mode: Access Point
Errors: OK for my config
This is the first public build that works for wdr4900v2 in terms of 5Ghz.
dnsmasq is hanging after about 5 minutes.
I couldn't do nslookup, or dig from the routers entware package, and when it would work it would constantly say "Truncated; trying tcp isntead."
After a lot of reading, and looking at the mailing list it appears when dnssec is enabled it is sending too large of packets over udp.
I use dnscrupt-proxy2 and have dnsmasq use that as the resolver address. dnsmasq runs great if I use our ISP's standard unicast addresses, but if I chain these 2 programs together it slows to a crawl and eventually stops responding.
If I use just dnsmasq alone, it works fine except for the horrible cache issue where it keeps hammering upstream servers looking for valid dnssec replies and invalidating the cache. If I just use dnscrypt-prox2 it also works fine, but pointing dnsmasq at dnscrypt-proxy2 local binding and problem starts after a few mins of operation.
I just turned off dnssec for dnsmasq and I still get the truncated errors, but it works much faster and now it caches everything but of course it isn't secure.
I may have to wait for the next version to re-enable dnssec.
Anyone else ran into this problem?
Can someone help me remember how to bind over the internal dnsmasq binary with the one I have in opt?
If I remember right (my memory is declining), I use stopservice dnsmasq, then use bind to bind over the binary is /usr/sbin/dnsmasq , then startservice dnsmasq and it should just retart with the same config but using a newer binary?
Last edited by roadrun777 on Fri Oct 11, 2019 4:34; edited 1 time in total
Upgraded from an old build that was released in March 2019 and so far everything seems fine. This router gets very little use per day unless it's a holiday and even then it's pretty minimal. One question I had is if anyone knew if there was a reason this router doesn't have the Cubic TCP congestion control algorithm. My E3200 (Broadcom) is running the same kernel and it has Cubic but I noticed this router doesn't. It's no big deal I just wondered if there was a reason.