Hi Todd, Last 4 days I'm reading all threads with sks tunning, suggestions / troubles / errors etc.. It's really large dose of valuable informations. SKS gives me different surprises every day. Final tunning its one of most important thing but it's very difficult to prevent malfunctions with unknown reasons. Today new one again - SKS shutting down every few hours. When restarting manually, situation repeats itself after few hours.
sks.service - SKS database service Loaded: loaded (/lib/systemd/system/sks.service; enabled; vendor preset: disabled) Active: failed (Result: signal) since Sat 2020-10-17 15:55:14 CEST; 22min ago Docs: man:sks(8) Process: 3474 ExecStart=/usr/sbin/sks -stdoutlog db (code=killed, signal=SEGV) Main PID: 3474 (code=killed, signal=SEGV) Oct 17 15:55:14 Khaos systemd[1]: sks.service: Main process exited, code=killed, status=11/SEGV Oct 17 15:55:14 Khaos systemd[1]: sks.service: Failed with result 'signal'. Keeping keys on servers like sks is a very important goal for a positive idea but costs much work and attention. br Adam ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐ piątek, 16 października 2020 18:33, Todd Fleisher <t...@fleetstreetops.com> napisał(a): > Adam, > You can also search the list archives for a thread with subject "SKS scaling > configuration” which goes into detail about how to build a more robust pool > of nodes to service requests. The software is far from perfect and you will > likely see some errors even under “normal” operation. I wouldn’t worry about > them unless they are causing specific issues. > > Marcin, > Very cool info on your website, thanks for sharing. > > -T > >> On Oct 16, 2020, at 06:00, Adam Wojcieszonek <a...@eksploracja.org.pl> wrote: >> >> Hi, >> Thanks for valuable suggestion. Today night will try with Varnish cache. >> >> Dziękuję ,pozdrawiam serdecznie >> >> Adam >> >> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐ >> piątek, 16 października 2020 12:41, Marcin Gondek <drix...@e-utp.net> >> napisał(a): >> >>> Hi, >>> >>> https://fido.e-utp.net/display/EUTPNET/SKS+Server+Caching >>> >>> Maybe my old notes with dual SKS will help. >>> >>> Thanks, >>> >>> -- >>> >>> Marcin Gondek / Drixter >>> http://fido.e-utp.net/ >>> AS56662 >>> >>> --------------------------------------------------------------- >>> >>> Od: Sks-devel <sks-devel-bounces+drixter=e-utp....@nongnu.org> w imieniu >>> użytkownika Adam Wojcieszonek <a...@eksploracja.org.pl> >>> Wysłane: piątek, 16 października 2020 12:37 >>> Do: sks-devel@nongnu.org <sks-devel@nongnu.org> >>> Temat: Re: seeking peers for hyperboria.net.pl >>> >>> Hi again >>> My server (Debian 9) is configured according to mrjones plip blog >>> https://blog.plip.com/2018/06/29/deploying-a-pgp-sks-server-on-ubuntu-18-04/ >>> Looks like similar to other configuration tutorials but as i observe my >>> proxy is hanging every time. This causes srv is thrown every hour from the >>> pool. Does anyone have an idea how to fix it ?? >>> I've testet adding "retry=0" to web proxy configuration and also extend >>> timeouts in apache2.conf by adding "Timeout 2400, ProxyTimeout 2400, >>> ProxyBadHeader Ignore" but nothing changes. Can You give some examples of >>> Apache configuration ? (tried also to search Google and this mailing list >>> but no right fixes found for SKS). >>> I am also worried about errors in the log that I wrote about yesterday >>> night. >>> >>> br >>> >>> Adam >>> >>> Sks running few hours and I already have few questions. >>> I have traced syslog and can see frequently recurring event logs. Not sure >>> something is wrong with sksconf ? >>> >>> 1. >>> >>> Oct 16 00:58:32 Khaos sks[10527]: 2020-10-16 00:58:32 99 keys received >>> Oct 16 00:59:32 Khaos sks[10526]: 2020-10-16 00:59:32 add_keys_merge >>> failed: Eventloop.SigAlarm >>> Oct 16 00:59:32 Khaos sks[10526]: 2020-10-16 00:59:32 Key addition failed: >>> Eventloop.SigAlarm >>> >>> (last few hours see 0 updated keys in stats page but DB folder size growing >>> really fast . After Eventloop.SigAlarm sks instance is unresponsive few >>> minutes and cannot enter stats page) >>> >>> 2. >>> Oct 16 00:52:11 Khaos sks[10526]: >>> x-forwarded-server:keyserver.hyperboria.net.pl]): Sys_error("Connection >>> reset by peer") >>> >>> 3. >>> Oct 16 00:50:00 Khaos sks[771]: host:127.0.0.1:11371 >>> Oct 16 00:50:00 Khaos sks[771]: pragma:no-cache >>> Oct 16 00:50:00 Khaos sks[771]: via:1.1 keyserver.hyperboria.net.pl:11371 >>> Oct 16 00:50:00 Khaos sks[771]: x-forwarded-for:217.76.45.34 >>> Oct 16 00:50:00 Khaos sks[771]: >>> x-forwarded-host:pool.sks-keyservers.net:11371 >>> Oct 16 00:50:00 Khaos sks[771]: >>> x-forwarded-server:keyserver.hyperboria.net.pl]): Sys_error("Broken pipe") >>> >>> Can someone explain me what above does it mean ? >>> >>> Here is conf with addressess. IP's should be local , external IP or leave >>> as it is 127.0.0.1 ? >>> # recon_address: 127.0.0.1 >>> recon_port: 11370 >>> hkp_address: 127.0.0.1 ::1 >>> hkp_port: 11371 >>> >>> Adam
