pmm-server stops collect localhost metrics after reboot

[DmytroKh]

Hi,
I’ve installed pmm server from image on amazon instance and upgraded it to v.1.5.3.
from first it looks OK, but than I reboot the server to upgrade a linux kernel.

I’ve noticed the issue that [url]http://localhost:9100/metrics[/url] is DOWN since the reboot (it used to be up before)

of course nothing listens the port 9100 on pmm-server while all configs are on place and prometheus is running.

I assume some exporter should be running on 9100 to provide linux metrics data, how can it be restored ?

[Mykola]

Hi Dmytro,

what kind of image do you use? Docker or AMI?

[fredricj]

I have the same problem. After updating the server the data is coming in fine from the server, but if I restart the pmm-server the client doesn’t run. Running pmm-admin list tells me the client isnt configured. I used the OVF image for the server

[DmytroKh]

HI,

the issue observed with AMI image, I switched to Docker and do not see the issue any more.

[otooi]

It seems to be related to the service startup sequence.
I solved the problem as shown below.

---

systemctl stop pmm-managed.service;
systemctl stop pmm-manage.service;
systemctl stop prometheus.service;
systemctl stop prometheus1.service;
systemctl stop percona-qan-api.service;
systemctl stop node_exporter.service;
systemctl stop pmm-qan-agent-9000.service;
systemctl stop pmm-rds_exporter-9042.service;

Stop all pmm-mysqld_exporter - *. Service. They will be automatically generated later.

systemctl stop pmm-mysqld_exporter-10001.service;
systemctl stop pmm-mysqld_exporter-10002.service;
systemctl stop nginx.service;
systemctl stop grafana-server.service;

ls /etc/systemd/system/pmm-;
rm -f /etc/systemd/system/pmm-;

systemctl start prometheus.service;
systemctl start prometheus1.service;
systemctl start pmm-manage.service;
systemctl start nginx.service;
systemctl start grafana-server.service;
systemctl start percona-qan-api.service;
systemctl start pmm-managed.service;