Question

我正在尝试使用一些自定义资源代理（由Bitnine https://github.com/bitnine-oss编写）为HA Postgres-XL使用Pacemaker资源。

我的过程是： 1.创建原始资源“ data1”，其中包括启动/停止/升级/删除功能。 2.创建控制原语实例的主/从资源。

在这里，我希望在节点t2上启动主节点，在节点t4上启动从节点。我正在尝试使用以下位置评分来控制主/从资源的位置。

@ubuntu:~$ sudo pcs -f data1.xml create resource data1 ocf:bitnine:postgres-xl-data datadir=/var/lib/postgresql/DATA/pgxl/nodes/data1 port=15432 nodename=data1
@ubuntu:~$ sudo pcs -f data1.xml resource master data1-ha data1 master-max=1 master-node-max=1 clone-max=2 clone-node-max=1 target-role="Started" notify="True"
@ubuntu:~$ sudo pcs constraint location data1-ha prefers t2=1000 t4=10

但是，出现以下错误。因为我的代理人描述性不是很高，所以我通过实验发现它试图将两个节点都作为主节点而不是主节点/从节点启动。

@ubuntu:~$ sudo pcs status
[sudo] password for michaelscott: 
Cluster name: xl-cluster

WARNINGS:
No stonith devices and stonith-enabled is not false

Stack: corosync
Current DC: t4 (version 1.1.18-2b07d5c5a9) - partition with quorum
Last updated: Thu Mar 28 18:19:32 2019
Last change: Thu Mar 28 17:48:48 2019 by root via crm_resource on t1

5 nodes configured
2 resources configured

Online: [ t1 t2 t3 t4 t5 ]

Full list of resources:

 Master/Slave Set: data1-ha [data1]
     data1  (ocf::bitnine:postgres-xl-data):    FAILED Master t4
     Masters: [ t2 ]
     Stopped: [ t1 t3 t5 ]

Failed Actions:
* data1_monitor_11000 on t4 'master' (8): call=22610, status=complete, exitreason='',
    last-rc-change='Thu Mar 28 18:19:32 2019', queued=0ms, exec=107ms


Daemon Status:
  corosync: active/disabled
  pacemaker: active/disabled
  pcsd: active/enabled

如果我更改“ master-max = 2”，则会得到以下信息：

@ubuntu:~$ sudo pcs status
[sudo] password for michaelscott: 
Cluster name: xl-cluster

WARNINGS:
No stonith devices and stonith-enabled is not false

Stack: corosync
Current DC: t4 (version 1.1.18-2b07d5c5a9) - partition with quorum
Last updated: Thu Mar 28 18:19:32 2019
Last change: Thu Mar 28 17:48:48 2019 by root via crm_resource on t1

5 nodes configured
2 resources configured

Online: [ t1 t2 t3 t4 t5 ]

Full list of resources:

 Master/Slave Set: data1-ha [data1]
     data1  (ocf::bitnine:postgres-xl-data):    
     Masters: [ t2 t4 ]


Daemon Status:
  corosync: active/disabled
  pacemaker: active/disabled
  pcsd: active/enabled

我的问题是：如何强制在t2上启动Master实例，在t4上启动Slave？我花了几天的时间来阅读有关Pacemaker中位置/共置的文档/示例（我很直觉可以找到答案），但是到目前为止还没有任何东西可以工作。

谢谢

编辑：

今天，进一步尝试通过将约束添加到CIB中来限制位置，以将母版推到t2，如下所示。同样的错误。

  <constraints>
      <rsc_location id="location-data1-ha-t1--INFINITY" node="t1" rsc="data1-ha" score="-INFINITY"/>
      <rsc_location id="location-data1-ha-t3--INFINITY" node="t3" rsc="data1-ha" score="-INFINITY"/>
      <rsc_location id="location-data1-ha-t5--INFINITY" node="t5" rsc="data1-ha" score="-INFINITY"/>
      <rsc_location id="location-data1-ha" rsc="data1-ha">
        <rule id="location-data1-ha-rule" role="master" score="INFINITY">
          <expression attribute="#uname" id="location-data1-ha-rule-expr" operation="eq" value="t2"/>
        </rule>
      </rsc_location>
      <rsc_location id="location-data1-ha-1" rsc="data1-ha">
        <rule id="location-data1-ha-1-rule" role="slave" score="INFINITY">
          <expression attribute="#uname" id="location-data1-ha-1-rule-expr" operation="eq" value="t4"/>
        </rule>
      </rsc_location>
    </constraints>

用于PostreSQL的Pacemaker主/从资源

0 个答案: