docker-library / healthcheck Goto Github PK

Thanks for providing this useful container! I have a question about two cassandra configs: enable_materialized_views and enable_transient_replication. It seems the official document
does not recommend to enable them in production but but they are set to be enabled in /etc/cassandra/cassandra.yaml:

Shall we disable them?
Thanks!

Maybe this isn't the proper place to ask, but is there any particular reason the prototyping that has occurred within these images haven't made its way into the official images? HEALTHCHECK is an advertised feature of Dockerfiles and it feels that official images should show the way forward and demonstrate how to check upon themselves.

When container is in multiple networks, most of the healthchecks fail.
The reason is that commands like

host="$(hostname -i || echo '127.0.0.1')"

(from postgres, redis), and

host="$(hostname --ip-address || echo '127.0.0.1')"

(from elasticsearch)

use hostname -i which in case of multiple networks return multiple ips.
i.e.

[root@ccaa85c77ed1 elasticsearch]# hostname --ip-address
192.168.208.3 192.168.176.2

so $host becomes invalid.

Healthcheck scripts didn't work (timeout and service unhealthy in docker ps) in alpine images of rabbitmq and postgres, until i added the following command in Dockerfile:
RUN chmod +x /usr/local/bin/docker-healthcheck

Docker CE v17.03.0-ce-mac2 (15654)

To keep alpine images as light and as close to as possible to upstream, the healthcheck scripts here could be for plain /bin/sh instead.

Hello all!

Thanks for sharing this. I've tried your healtcheck method with postgres using https://hub.docker.com/r/mdillon/postgis/~/dockerfile/ as the base image and everything works as expected when in non-swarm mode. Whenever I try to the same image in swarm mode, it never stays in running mode. Is it incompatible? It looks like it is ready to run, but, after a while, there is a shutdown request, as seen in the logs:

test-swarm_db.1.w83nr7m4m3wq@test-docker-swarm-test2-manager    | LOG:  received smart shutdown request
test-swarm_db.1.w83nr7m4m3wq@test-docker-swarm-test2-manager    | LOG:  autovacuum launcher shutting down
test-swarm_db.1.w83nr7m4m3wq@test-docker-swarm-test2-manager    | LOG:  shutting down
test-swarm_db.1.w83nr7m4m3wq@test-docker-swarm-test2-manager    | LOG:  database system is shut down
test-swarm_db.1.juy14y8pi8xu@test-docker-swarm-test2-manager    | LOG:  database system was shut down at 2017-03-21 16:43:58 UTC

Thank you all in advance!

EDIT: Nvm, it looks like everything is ok - I just add a versioning problem....

Healthcheck command fails if redis.conf is provided. I ran this manually and I get this error:

Could not connect to Redis at 172.17.0.2:6379: Connection refused

My assumption is either localhost or 127.0.0.1 should stay, instead of fetching the host ip

Here is my Dockerfile

Thanks for all the work you're doing @tianon 👍
Would love to see alpine images with health check :) postgres:latest is 100mb+ while postgres:alpine is 14mb

Hi,
you can try using mysqladmin to check whether the MySQL server is alive. I do the following in a project for blocking the docker start sequence until the server has been initialized. What happens is that a COM_PING packet is sent over the wire to the server. Here is some documentation about the COM_PING packet:
https://dev.mysql.com/doc/internals/en/com-ping.html

                while [[ `mysqladmin ping -uroot -p$$MYSQL_ROOT_PASSWORD 2>/dev/null` != "mysqld is alive" ]] ; \
                        do echo Waiting MySQL Server to start...; sleep 1; \
                done ; \

A SELECT query will increase a statistical variable about the queries. If someone is using it this will skew the results.

It would be nice to be able to use specific version of postgres:

for instance:

postgres:
image: healthcheck/postgres:9.5

Hi!

I am using the healthcheck/mysql. The interval for healthcheck seems to be set to 30s by default from my observation. I am wondering is there a way to change the interval?

Thanks!!