[SOLVED] Serious mupx/Docker issue - help!

deployment
1

I’m in a huge bind. This is the second time that this is occurring, where the Docker status is “Restarting” for a mission critical app.

CONTAINER ID        IMAGE                      COMMAND                  CREATED             STATUS                       PORTS                        NAMES
febfd05b2e48        mongo                      "/entrypoint.sh mongo"   2 minutes ago       Up 2 minutes                 127.0.0.1:27017->27017/tcp   mongodb
40e3072e7d9f        meteorhacks/meteord:base   "/bin/sh -c 'bash $ME"   17 minutes ago      Up 17 minutes                0.0.0.0:3002->80/tcp         app1
d38dc836541a        meteorhacks/meteord:base   "/bin/sh -c 'bash $ME"   46 hours ago        Restarting (8) 4 hours ago   0.0.0.0:3000->80/tcp         app2

docker logs app2 reveals:

Error: failed to connect to [mongodb:27017]

But localhost port 27017 is fine, and in fact app1 has no problems. Why is this happening??

2

@arunoda - any ideas?

3

I take that back about app1 working. I rebooted the server, and now both apps are failing:

CONTAINER ID        IMAGE                      COMMAND                  CREATED             STATUS                              PORTS                        NAMES
febfd05b2e48        mongo                      "/entrypoint.sh mongo"   15 minutes ago      Up 3 minutes                        127.0.0.1:27017->27017/tcp   mongodb
40e3072e7d9f        meteorhacks/meteord:base   "/bin/sh -c 'bash $ME"   30 minutes ago      Restarting (8) About a minute ago   0.0.0.0:3002->80/tcp         app1
d38dc836541a        meteorhacks/meteord:base   "/bin/sh -c 'bash $ME"   46 hours ago        Restarting (8) About a minute ago   0.0.0.0:3000->80/tcp         app2

Same error, can’t connect to mongodb. I can run the mongo client just fine and connect to the DB, run queries, etc.

4

I’m narrowing in on the issue. The link that should be there is broken. If I do docker inspect app1 I see this:

    "Links": null,

So the mongodb container link is not there as it should be.

5

Solved. I’ll leave this thread up, as a reminder: always rm -rf ~/.meteor if things are acting oddly!

The clue was the first few lines in the mupx deploy errors:

/var/lib/jenkins/.meteor/packages/meteor-tool/.1.1.10.aci3f7++os.linux.x86_64+web.browser+web.cordova/mt-os.linux.x86_64/dev_bundle/lib/node_modules/fibers/future.js:245
						throw(ex);
						      ^
Error: ENOSPC, open '/tmp/eea7cd97-7373-4352-825f-fd729440be65/.build267133.bundle/programs/.build561573.server/npm/babel-compiler/node_modules/meteor-babel/node_modules/iconv-lite/Changelog.md'
    at Object.Future.wait (/var/lib/jenkins/.meteor/packages/meteor-tool/.1.1.10.aci3f7++os.linux.x86_64+web.browser+web.cordova/mt-os.linux.x86_64/dev_bundle/lib/node_modules/fibers/future.js:398:15)

It seemed odd that I was getting errors about Meteor packages, and particularly ENOSPC even though I had plenty of disk space (still not sure why node would say that).

At any rate, rm -rf ~/.meteor did the trick and everything is fine now!

1 Like
6

At any rate, rm -rf ~/.meteor did the trick and everything is fine now!

Thanks man! :thumbsup:

Also I’ve freed a huge amount of disk space! :wink:

1 Like