如果一项服务失败,如何配置 crm 迁移 ServiceIP?
node 1: web01a \
attributes standby=off
node 2: web01b \
attributes standby=off
primitive Apache2 systemd:apache2 \
operations $id=Apache2-operations \
op start interval=0 timeout=100 \
op stop interval=0 timeout=100 \
op monitor interval=15 timeout=100 start-delay=15 \
meta
primitive PHP-FPM systemd:php7.4-fpm \
operations $id=PHP-FPM-operations \
op start interval=0 timeout=100 \
op stop interval=0 timeout=100 \
op monitor interval=15 timeout=100 start-delay=15 \
meta
primitive Redis systemd:redis-server \
operations $id=Redis-operations \
op start interval=0 timeout=100 \
op stop interval=0 timeout=100 \
op monitor interval=15 timeout=100 start-delay=15 \
meta
primitive ServiceIP IPaddr2 \
params ip=1.2.3.4 \
operations $id=ServiceIP-operations \
op monitor interval=10 timeout=20 start-delay=0 \
op_params migration-threshold=1 \
meta
primitive lsyncd systemd:lsyncd \
op start interval=0 timeout=100 \
op stop interval=0 timeout=100 \
op monitor interval=15 timeout=100 start-delay=15 \
meta target-role=Started
group ActiveNode ServiceIP lsyncd
group WebServer Apache2 PHP-FPM Redis
clone cl_WS WebServer \
meta clone-max=2 notify=true interleave=true
colocation col_cl_WS_ActiveNode 100: cl_WS ActiveNode
property cib-bootstrap-options: \
have-watchdog=false \
dc-version=2.0.3-4b1f869f0f \
cluster-infrastructure=corosync \
cluster-name=debian \
stonith-enabled=false \
no-quorum-policy=ignore \
startup-fencing=false \
maintenance-mode=false \
last-lrm-refresh=1622628525 \
start-failure-is-fatal=true
这些服务应该始终启动
- 阿帕奇2
- PHP-FPM
- 雷迪斯
如果其中一项服务未运行,则该节点不健康。这服务IP和同步应该切换到健康的节点。
当我杀死apache2进程时,IP没有切换。