I just had a "freeze" on r37985 after a week of uptime. Everything just stopped working, including the router GUI. After a reboot, all is fine. Nobody is even here and I was watching antenna TV - traffic would have been nearly nothing so it wouldn't have been a load issue.
37985 is crap i wouldn't trust it. use either 38065 or the latest 38100
I'm about 2 weeks behind on updates. It seems like about the time I get around to updating, Kong is releasing a new version so I was always about a week behind. Then the holidays came and I haven't even been on a computer since December 21st.
I'm on r38100 now and we'll see how it goes. I haven't had any problems with any DD-WRT builds until today. _________________ Routing:.......Asus RT-AX88U (Asuswrt-Merlin 384.14) Switching:....Netgear GS608_V3 & GS605_V4, TrendNet TEG-S82G & TEG-S50G
Had to boot back into other part r38065M kongat 12/22/18
I was still able to ssh into r38100M but it had quit resolving and showed over 3800 active IP connections.
Code:
Dec 27 06:46:00 --- daemon.warn dnsmasq[1752]: Maximum number of concurrent DNS queries reached (max: 150)
Dec 27 06:46:20 --- daemon.warn dnsmasq[1752]: Maximum number of concurrent DNS queries reached (max: 150)
Dec 27 06:46:34 --- daemon.warn dnsmasq[1752]: Maximum number of concurrent DNS queries reached (max: 150)
Just using QUAD9 DNS + local DNS
Same as some previous builds but r38065M never gave trouble.
mrjcd, about your recent dns issue, if you use static dns, I found removing the isp entries in the resolver works out well for me, resolves faster and much stable, especially with ones that support dnssec. Mine is cloudflare with 2 entries 1.1.1.1/1.0.0.1 hence 2 in the startup script below:
You can do cat /tmp/resolv.dnsmasq to see if your isp entries are in there before and after startup script/reboot
p.s. believe it or not i used to have the exact same issue as well until i used that startup command. before that i also tried increasing the cache size to 1000 but didn't work so no longer needed
mrjcd, about your recent dns issue, if you use static dns, I found removing the isp entries in the resolver works out well for me, resolves faster and much stable, especially with ones that support dnssec. Mine is cloudflare with 2 entries 1.1.1.1/1.0.0.1 hence 2 in the startup script below:
You can do cat /tmp/resolv.dnsmasq to see if your isp entries are in there before and after startup script/reboot
p.s. believe it or not i used to have the exact same issue as well until i used that startup command. before that i also tried increasing the cache size to 1000 but didn't work so no longer needed
I have NO problems with r38065M kongat 12/22/18
and QUAD9 is fully DNSSEC. I check DNS few times a day from different places and QUAD9 is all that is used...... I might turn unbound back on someday but I don't really even see the point
Thank you. I read a post in the Broadcom forum a while back by <Kong> that nvram erase was correct for builds after April or so. I wasn't sure if that applied only to Broadcom. Was not sure <Kong> builds and STD DD-WRT builds used the same either. Also was not trusting the wiki to be up to date It might not hurt for wiki pages to show update dates.
Joined: 03 Jan 2010 Posts: 7568 Location: YWG, Canada
Posted: Sun Dec 30, 2018 20:18 Post subject:
dnsmasq just crashed from OOM though the router had over 60% free ram..it seemed to have restarted itself right away so i didnt notice till i check syslog.
have 564k lines in adblock.txt. dnsmasq creator says its been tested to 1mill lines (cough, say that again for many other qca routers that still cant run over 350 lines). my only other dnsmasq config settings are
dnsmasq just crashed from OOM though the router had over 60% free ram..it seemed to have restarted itself right away so i didnt notice till i check syslog.
have 564k lines in adblock.txt. dnsmasq creator says its been tested to 1mill lines (cough, say that again for many other qca routers that still cant run over 350 lines). my only other dnsmasq config settings are
Thank you. I read a post in the Broadcom forum a while back by <Kong> that nvram erase was correct for builds after April or so. I wasn't sure if that applied only to Broadcom. Was not sure <Kong> builds and STD DD-WRT builds used the same either. Also was not trusting the wiki to be up to date It might not hurt for wiki pages to show update dates.
Never mind that last statement, I just found it at the bottom = History. Thanks Again!
Sorry I can't find the post, but I recall
<Kong> stated that he tested the erase nvram command to work properly on all the routers that he supports and that he could not say the same for the ones BrainSlayer does
Obviously that was a while ago though as the command is now reversed..nor do I know if it still stands true
Not necessarily related but I've run the old command on my WNDR3700v2 (Atheros) without issue and it performed as expected in the past _________________ Location 1
R7800- DD-WRT v3.0-r53562 (10/03/23) Gateway
WNDR3400v1 DD-WRT v3.0-r35531_mega-nv64k (03/26/18 ) Access Point
WRT160Nv3 DD-WRT ?v3?.0-r35531 mini (03/26/18 ) Access Point
WRT54GSv5 DD-WRT v24-r33555_micro_generic (10/20/17) Repeater
Location 2
R7800- DD-WRT v3.0-r51855 (02/25/23) Gateway
R6300v2- DD-WRT v3.0-r50671 (10-26-22) Access Point
WNDR3700v2 DD-WRT v3.0-r35531 std (03/26/18 ) Access Point
E1200 v2 DD-WRT v3.0-r35531 mega-nv64k (03/26/18 ) Gateway(for trivial reasons)
RBWAPG-5HACT2HND-BE RouterOS-v6.46.4 (2/21/20) Outdoor Access Point
2x RBSXTG-5HPACD RouterOS-v6.46.4 (2/21/20) PTP Bridge 866.6Mbps-1GbpsLAN
Location 3
2x R7000- DD-WRT v3.0-r50671 (10/26/22) Access Points
2x RBWAPG-60AD RouterOS-v6.45.9 (04/30/20) PTP Bridge 2.3Gbps-1GbpsLAN
2x RBSXTsqG-5acD RouterOS-v6.49.7 (10/14/22) PTP Bridge 866.6Mbps-1GbpsLAN Thank You BrainSlayer for ALL that you do & have done, also to "most" everyone here that shares their knowledge
included tickets are only those affecting r7800 and kong build only.
wifi load still isnt normal and there is elevated LOCAL latency to the router from extra 5ms~15ms. not a single wifi setting effects it in the slightest.
I just tested pings, I can reproduce it on my ipda, but intel ac 9560 works fine. Thus not a general issue, but likely a compatibility thing. _________________ KONG PB's: http://www.desipro.de/ddwrt/
KONG Info: http://tips.desipro.de/
Thank you. I read a post in the Broadcom forum a while back by <Kong> that nvram erase was correct for builds after April or so. I wasn't sure if that applied only to Broadcom. Was not sure <Kong> builds and STD DD-WRT builds used the same either. Also was not trusting the wiki to be up to date It might not hurt for wiki pages to show update dates.
Never mind that last statement, I just found it at the bottom = History. Thanks Again!
Sorry I can't find the post, but I recall
<Kong> stated that he tested the erase nvram command to work properly on all the routers that he supports and that he could not say the same for the ones BrainSlayer does
Obviously that was a while ago though as the command is now reversed..nor do I know if it still stands true
Not necessarily related but I've run the old command on my WNDR3700v2 (Atheros) without issue and it performed as expected in the past
No the command is not reversed, but both do different things, only one thing changed, that causes issues if you think, now if you reboot nvram cache from ram is flushed and thir vars are written to nvram flashspace. This will write parts of nvram to flash again.
Thus erasen nvram won't wipe everything unless you press the power button right afterwards. nvram erase clears nvram space, but does not zero out the flash from first to last block. _________________ KONG PB's: http://www.desipro.de/ddwrt/
KONG Info: http://tips.desipro.de/
No the command is not reversed, but both do different things, only one thing changed, that causes issues if you think, now if you reboot nvram cache from ram is flushed and thir vars are written to nvram flashspace. This will write parts of nvram to flash again.
Thus erasen nvram won't wipe everything unless you press the power button right afterwards. nvram erase clears nvram space, but does not zero out the flash from first to last block.
so which do you recommend? erase nvram + power cycle the rt?
included tickets are only those affecting r7800 and kong build only.
wifi load still isnt normal and there is elevated LOCAL latency to the router from extra 5ms~15ms. not a single wifi setting effects it in the slightest.
I just tested pings, I can reproduce it on my ipda, but intel ac 9560 works fine. Thus not a general issue, but likely a compatibility thing.
its affecting every single device thats not a wds station. iphones, androids, intel, broadcom.. been happening since k4.9 switch
all same devices work fine on qca9880. check same device again on k3.18, fine right? and any other ac router u have, fine again right? _________________ LATEST FIRMWARE(S)
BrainSlayer wrote:
we just do it since we do not like any restrictions enforced by stupid cocaine snorting managers
No the command is not reversed, but both do different things, only one thing changed, that causes issues if you think, now if you reboot nvram cache from ram is flushed and thir vars are written to nvram flashspace. This will write parts of nvram to flash again.
Thus erasen nvram won't wipe everything unless you press the power button right afterwards. nvram erase clears nvram space, but does not zero out the flash from first to last block.
so which do you recommend? erase nvram + power cycle the rt?
I second that question...as most of us take your recommendations as golden rules
<Kong> Thank You for chiming in to help clear this up for us.....And for all that you do here _________________ Location 1
R7800- DD-WRT v3.0-r53562 (10/03/23) Gateway
WNDR3400v1 DD-WRT v3.0-r35531_mega-nv64k (03/26/18 ) Access Point
WRT160Nv3 DD-WRT ?v3?.0-r35531 mini (03/26/18 ) Access Point
WRT54GSv5 DD-WRT v24-r33555_micro_generic (10/20/17) Repeater
Location 2
R7800- DD-WRT v3.0-r51855 (02/25/23) Gateway
R6300v2- DD-WRT v3.0-r50671 (10-26-22) Access Point
WNDR3700v2 DD-WRT v3.0-r35531 std (03/26/18 ) Access Point
E1200 v2 DD-WRT v3.0-r35531 mega-nv64k (03/26/18 ) Gateway(for trivial reasons)
RBWAPG-5HACT2HND-BE RouterOS-v6.46.4 (2/21/20) Outdoor Access Point
2x RBSXTG-5HPACD RouterOS-v6.46.4 (2/21/20) PTP Bridge 866.6Mbps-1GbpsLAN
Location 3
2x R7000- DD-WRT v3.0-r50671 (10/26/22) Access Points
2x RBWAPG-60AD RouterOS-v6.45.9 (04/30/20) PTP Bridge 2.3Gbps-1GbpsLAN
2x RBSXTsqG-5acD RouterOS-v6.49.7 (10/14/22) PTP Bridge 866.6Mbps-1GbpsLAN Thank You BrainSlayer for ALL that you do & have done, also to "most" everyone here that shares their knowledge
Joined: 03 Jan 2010 Posts: 7568 Location: YWG, Canada
Posted: Mon Dec 31, 2018 1:50 Post subject:
<Kong> wrote:
tatsuya46 wrote:
dnsmasq just crashed from OOM though the router had over 60% free ram..it seemed to have restarted itself right away so i didnt notice till i check syslog.
have 564k lines in adblock.txt. dnsmasq creator says its been tested to 1mill lines (cough, say that again for many other qca routers that still cant run over 350 lines). my only other dnsmasq config settings are
Will merge my fixes for dnsmasq from northstar build, then you can check if you still have problems.
ive tested a adblock file even bigger than current and it ran fine.. do u want mine to test with? if i had to guess i think maybe its related to the 10000 cache entries getting filed, i would think it was up long enough to do so, about 1.5 days _________________ LATEST FIRMWARE(S)
BrainSlayer wrote:
we just do it since we do not like any restrictions enforced by stupid cocaine snorting managers