[TABLE]
[TR]
[TD]I monitored multiple mysql instances on one server:
Pmm-admin add mysql --user PMM --password PMM --socket= XXX 1.1.1.1-3308
Pmm-admin add mysql --user PMM --password PMM --socket= XXX 1.1.1.1-3309
Pmm-admin add mysql --user PMM --password PMM --socket= XXX 1.1.1.1-3310
And add Linux :metric
PMM - admin add Linux: metric 1.1.1.1
Now there is no data in the buffer pool size of total ram in the mysql overview of the monitoring interface. Does the mysql instance 1.1.1.1-3308 need to read the data of Linux :metric 1.1.1.1-3308? But what I added is the data of Linux :metric 1.1.1.1?
I tried to add more Linux :metric X.X.X.X, but the system could only add one.How can you make buffer pool size of total ram of multiple instances display data normally[/TD]
[/TR]
[/TABLE]
Hi liangyu5017
Based on the way you added hosts you probably have 3 hosts in the MySQL dashboards, and one host in the OS dashboards.
PMM is limited to how many times you can use a friendly name, that is by default the hostname but you can override it. As an example I’ve added the same MySQL instance multiple times while changing the friendly name (ps1, ps2, ps3):
[root@ip-10-178-0-149 ~]# pmm-admin add mysql:metrics --socket=/tmp/PS_NODE_1.sock ps1
OK, now monitoring MySQL metrics using DSN root:***@unix(/tmp/PS_NODE_1.sock)
[root@ip-10-178-0-149 ~]# pmm-admin add mysql:metrics --socket=/tmp/PS_NODE_1.sock ps2
OK, now monitoring MySQL metrics using DSN root:***@unix(/tmp/PS_NODE_1.sock)
[root@ip-10-178-0-149 ~]# pmm-admin add mysql:metrics --socket=/tmp/PS_NODE_1.sock ps3
OK, now monitoring MySQL metrics using DSN root:***@unix(/tmp/PS_NODE_1.sock)
[root@ip-10-178-0-149 ~]# pmm-admin list
pmm-admin 1.17.0
PMM Server | 127.0.0.1
Client Name | ip-10-178-0-149
Client Address | 172.17.0.1
Service Manager | linux-upstart
-------------- ----- ----------- -------- ----------------------------------- --------
SERVICE TYPE NAME LOCAL PORT RUNNING DATA SOURCE OPTIONS
-------------- ----- ----------- -------- ----------------------------------- --------
mysql:metrics ps1 42002 YES root:***@unix(/tmp/PS_NODE_1.sock)
mysql:metrics ps2 42003 YES root:***@unix(/tmp/PS_NODE_1.sock)
mysql:metrics ps3 42004 YES root:***@unix(/tmp/PS_NODE_1.sock)
However with Linux metrics you can only add this service once per instance, even if you try to use a different friendly name:
[root@ip-10-178-0-149 ~]# pmm-admin add linux:metrics ps1
OK, now monitoring this system.
[root@ip-10-178-0-149 ~]# pmm-admin add linux:metrics ps2
Error adding linux metrics: there is already one instance with this name under monitoring.
[root@ip-10-178-0-149 ~]# pmm-admin list
pmm-admin 1.17.0
PMM Server | 127.0.0.1
Client Name | ip-10-178-0-149
Client Address | 172.17.0.1
Service Manager | linux-upstart
-------------- ----- ----------- -------- ----------------------------------- --------
SERVICE TYPE NAME LOCAL PORT RUNNING DATA SOURCE OPTIONS
-------------- ----- ----------- -------- ----------------------------------- --------
linux:metrics ps1 42000 YES -
mysql:metrics ps1 42002 YES root:***@unix(/tmp/PS_NODE_1.sock)
mysql:metrics ps2 42003 YES root:***@unix(/tmp/PS_NODE_1.sock)
mysql:metrics ps3 42004 YES root:***@unix(/tmp/PS_NODE_1.sock)
You can see InnoDB Buffer Pool Size when looking at the MySQL instances you added ,not the linux instance name.
I hope this helps - did I miss answering any questions?
Please look at the following two pictures:Master and slave are on the same host,As you can see, the master’s buffer size of total ram is normal,but the backup’s buffer size of total ram is no value.
Is this a bug? or Will you make improvements in the next version?
Hi liangyu5017
That looks more like a bug. Can you open the SingleStat object and go to Query Inspector? I think you’ll find an error, please share that with us so we can continue troubleshooting. Thanks!
