On 05/07/2018 11:02 AM, Aki Tuomi wrote: >>> Can you run sudo strace -econnect,bind rec_control reload-lua-script and >>> post the result? >> >> yes, thanks for your help >> >> bind(3, {sa_family=AF_LOCAL, sun_path="/var/run/lsock6hPxMw"}, 110) = 0 >> connect(3, {sa_family=AF_LOCAL, >> sun_path="/var/run/pdns_recursor.controlsocket"}, 110) = 0 >> Fatal: Timeout waiting for answer from control channel >> +++ exited with 1 +++ >> >> Yet in the log I still see the error >> pdns_recursor: 0 (Re)loaded lua script from >> '/etc/pdns-recursor/blocklist.lua' >> ... >> pdns_recursor: Error dealing with control socket request: Unable to >> send message over control channel '/var/run/lsock6hPxMw': No such file >> or directory >> >> Is the problem caused/related to the timeout?
What's happening is that the recursor is taking more than the default timeout of rec_control (5s) to do the actual reload, so rec_control stops waiting, delete its socket and exits. The recursor is still completing the reload, and tries to reply to the rec_control process over the socket afterwards, except the socket is now gone, hence the error. So yes, the reload is done as indicated by the "(Re)loaded lua script from ..." line in the logs. If you want to avoid the "Unable to send message over control channel" line in your logs, just use a longer timeout value for rec_control using the --timeout parameter. And please give us the relevant logs from the start next time, like including the "Fatal: Timeout waiting for answer from control channel" and "(Re)loaded lua script from" messages. You are much more likely to get help if we don't have to guess or ask. -- Remi Gacogne PowerDNS.COM BV - https://www.powerdns.com/
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Pdns-users mailing list Pdns-users@mailman.powerdns.com https://mailman.powerdns.com/mailman/listinfo/pdns-users