Infos:
Following the instructions at https://docs.zammad.org/en/latest/install-docker-compose.html I am trying to install zammad on a T2.medium instance on EC2. (T2.medium has 4GB RAM, so I hope its enough)
After invoking docker compose, it seems to be stuck in a loop:
zammad-elasticsearch_1 | [2018-03-29T03:50:47,399][INFO ][o.e.n.Node ] [4CqiwRh] closed
zammad-railsserver_1 | waiting for init container to finish install or update…
zammad-init_1 | zammad railsserver waiting for elasticsearch server to be ready…
zammad-websocket_1 | waiting for init container to finish install or update…
zammad-scheduler_1 | waiting for init container to finish install or update…
zammad-nginx_1 | waiting for init container to finish install or update…
zammad-backup_1 | waiting for railsserver to be ready…
zammad-init_1 | zammad railsserver waiting for elasticsearch server to be ready…
zammad-railsserver_1 | waiting for init container to finish install or update…
zammad-init_1 | zammad railsserver waiting for elasticsearch server to be ready…
zammad-websocket_1 | waiting for init container to finish install or update…
zammad-scheduler_1 | waiting for init container to finish install or update…
zammad-nginx_1 | waiting for init container to finish install or update…
zammad-init_1 | zammad railsserver waiting for elasticsearch server to be ready…
etc, then some other elasticsearch_1 stuff
zammad-elasticsearch_1 | ERROR: [1] bootstrap checks failed
zammad-elasticsearch_1 | [1]: max file descriptors [4096] for elasticsearch process is too low, increase to at least [65536]
zammad-elasticsearch_1 | [2018-03-29T03:54:43,130][INFO ][o.e.n.Node ] [4CqiwRh] stopping …
zammad-elasticsearch_1 | [2018-03-29T03:54:43,144][INFO ][o.e.n.Node ] [4CqiwRh] stopped
zammad-elasticsearch_1 | [2018-03-29T03:54:43,144][INFO ][o.e.n.Node ] [4CqiwRh] closing …
zammad-elasticsearch_1 | [2018-03-29T03:54:43,156][INFO ][o.e.n.Node ] [4CqiwRh] closed
Following https://github.com/docker-library/elasticsearch/issues/140 I added:
ulimits:
nofile:
soft: 65536
hard: 65536
which resolved the problem.