Hello,
We have a setup with hidden master (with couple of thousand zones) with
like 30 slaves which is slave of multiple BIND masters. The setup works
well but we are concerned about the subject.
If the first level master (where all the changes happen) is restarted
for example it sends notifys for all lets say 2k zones and the pdns
Master and mariadb start consume cpu but here is the strange thing
happen they share a single cpu core.
For example if the mariadb get 70% pdns 30% if i check the load by cores
one core has 0%id all other 3 are near 100%id .
here are some screenshots
https://monosnap.com/direct/tQ4YKdWEQNeZfwWrtXDYIGRNWaFeQB
https://monosnap.com/direct/h05o79glxSvobDTtgHqR5tLiwQLamm
The machine is KVM VM with 4 cores and 2GB ram with Centos 7
3.10.0-123.el7.x86_64
pdns 4.1.13
Mariadb 10.3.18
PDNS config
=================================
config-dir=/etc/powerdns
daemon=yes
disable-axfr=no
guardian=yes
local-ipv6=
local-address=0.0.0.0
local-port=53
loglevel=6
module-dir=/usr/lib64/pdns
master=no
slave=yes
slave-cycle-interval=60
setgid=pdns
setuid=pdns
socket-dir=/var/run
version-string=powerdns
launch=gmysql
gmysql-host=localhost
gmysql-user=XXXX
gmysql-dbname=XXXX
gmysql-password=XXXXXXXXXXXXXXXXX
#gmysql-socket=/var/lib/mysql/mysql.sock
slave-renotify=no
only-notify=
webserver=yes
webserver-address=XXX.XXX.XXX.XXX
webserver-allow-from=XXX.XXX.XXX.XXX
api=yes
api-key=YYYYYYYYYYYYYY
gmysql-dnssec=yes
retrieval-threads=4
receiver-threads=4
signing-threads=4
distributor-threads=4
max-tcp-connections=100
max-tcp-connection-duration=10
query-logging=no
disable-syslog=no
log-dns-details=no
reuseport=yes
==============================
If you need more information please let me know.
Warm regards
George A.
_______________________________________________
Pdns-users mailing list
Pdns-users@mailman.powerdns.com
https://mailman.powerdns.com/mailman/listinfo/pdns-users