Closed synchrone closed 5 years ago
Wow.
@silenius I'm going to need to roll your patch back until I figure out what is going on here.
@mbucc can you also trigger a rebuild in https://hub.docker.com/r/mkbucc/shmig/ ?
@synchrone Sure. Can you give me a test case so I can duplicate this bad behavior?
sure:
docker-compose.yml
version: '2.1'
services:
postgres:
image: kartoza/postgis
environment:
- ALLOW_IP_RANGE=0.0.0.0/0
ports: ['5452:5432']
migrations:
image: mkbucc/shmig:latest
restart: on-failure
volumes:
- ./migrations:/sql
environment:
- TYPE=postgresql
- HOST=postgres
- PORT=5432
- DATABASE=gis
- LOGIN=docker
- PASSWORD=docker
- UP_MARK=SHMIG
command: up
migrations/1509357658-init.sql
-- SHMIG
create extension if not exists postgis;
create type kind as enum ('type1');
create table mytable(
id int primary key
);
run docker-compose up
and after several restarts due to postgres starting up, it should hang for good.
FYI, I tried to duplicate the bug using your instructions way back when, but was not able to.
I cannot reproduce right now (on latest) either. I suppose this can be closed.
Docker container
mkbucc/shmig:latest
hangs in an endless loop, consuming 100% CPU (one core) after printingshmig: creating migrations table: shmig_version
.The only difference is https://github.com/mbucc/shmig/commit/0ef361297118f4c13c416eee244f268c28367259.
The issue is not present in
mkbucc/shmig:verbose
, which is currently built on2017-07-24T11:48:19.724Z
, and does not have this shebang patch applied.