4

I'm using sath89/oracle-12c for automated tests against a oracle db. This works fine, the only problem is that this container takes several minutes to start (~10-15 depending on the hardware). I tried to come up with a healthcheck for this container.

I managed to come up with

status=`su oracle -c "echo -e \"SELECT ACCOUNT_STATUS FROM DBA_USERS WHERE USERNAME = 'ANONYMOUS' AND ACCOUNT_STATUS = 'EXPIRED';\" | /u01/app/oracle/product/12.1.0/xe/bin/sqlplus -S / as sysdba | grep ACCOUNT_STATUS"`; if [ "$status" == "ACCOUNT_STATUS" ]; then true; else false; fi

which returns 0 when the ANONYMOUS account is unlocked, which is the last step in the entrypoint script of the image: entrypoint.sh. I tested this using docker exec -it <containername> bash.

I am now stuck with converting this horribly long line into a healthcheck command for docker (docker-compose):

version: "2" services: db: image: sath89/oracle-12c:r1 healthcheck: test: ["CMD", "<command goes here>"] interval: 10s timeout: 3s retries: 3

Any help is appreciated - if you can improve the command itself I'm happy to here. I am aware of "select 1 from dual" as a validation query for Oracle (source), but this reports an operational DB after ~8 minutes but it resets connections a little bit later. I don't want to modify the container itself - if there's an update I just want to be able to pull it from the hub.

Improve this question
3
  • Why don't you put the command into a bash script and add it to the image? Comming to your question, according to the docu, the command is executed via exec, thus you have to invoke bash and give it your script: test: ["CMD", "bash", "-c", "<horrible long script>"]. Commented Aug 31, 2017 at 14:12
  • I just realized you could also let docker-compose do the job of invoking the shell, though it uses /bin/sh: test: ["CMD-SHELL", "<horrible long script>"] Commented Aug 31, 2017 at 14:17
  • @fzgregor thanks for the hint to "CMD-SHELL" again. I stumbled across it myself but somehow ignored it. I have given it another try and it looks promising, will post an answer if it actually works. Commented Sep 5, 2017 at 12:48

1 Answer 1

Reset to default
6

Okay, so after quite some time I've come up with an solution for my problem. I could simplify the "" a bit:

version: '2.1' services: db: image: sath89/oracle-12c:r1 healthcheck: test: ["CMD-SHELL", "if [ \"`echo \\\"SELECT ACCOUNT_STATUS FROM DBA_USERS WHERE USERNAME = 'ANONYMOUS' AND ACCOUNT_STATUS = 'EXPIRED';\\\"|/u01/app/oracle/product/12.1.0/xe/bin/sqlplus -S sys/oracle as sysdba|grep ACCOUNT_STATUS`\" = \"ACCOUNT_STATUS\" ];then true;else false;fi"] interval: 30s timeout: 3s # start_period: 900s retries: 30

Right now "docker-compose" does not support the start_period option so the numbers of retries (and the interval) have to be quite high so the container isn't reported as "unhealthy". The Pull Request has already been merged so hopefully it will be in the next release.

Improve this answer
Sign up to request clarification or add additional context in comments.

1 Comment

your comment on the healthcheck retries is a bit off: retries defines the number of subsequent fails before a container becomes unhealthy. The healthcheck will in fact be run at every interval indefinitely. the check would return health: starting, after a while 'unhealthy', but after 'retries' number of checks returning 'ok', the container status will become healthy again.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.