fix(SSHChannelReader): Read output completely before checking command exit.

[calavera]

Fixes #259

[calavera]

FWIW, this switch will disappear in the future if this patch is accepted:

https://codereview.appspot.com/81290043/

[bcwaldon]

@calavera This still does not fix the problem. The issue is in runRemoteCommand [0]. The ssh session is being closed before readSSHChannel even has a chance to read its stdout buffer.

[0]

fleet/fleetctl/journal.go

Line 106 in 275ce5a

defer sshClient.Close()

[calavera]

oh, I didn't see that. I don't think that's necessary. Let me run some tests.

[calavera]

I think it's working as expected now.

This is the output in the server:

core@ip-10-197-133-18 ~ $ journalctl --no-pager -l -u subgun-http.1.service
-- Logs begin at Sun 2014-03-23 21:52:09 UTC, end at Wed 2014-04-02 02:22:11 UTC. --
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:8] Unknown section 'X-Fleet'. Ignoring.
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:9] Assignment outside of section. Ignoring.
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: Starting subgun...
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: Started subgun.
Apr 02 02:22:03 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:8] Unknown section 'X-Fleet'. Ignoring.
Apr 02 02:22:03 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:9] Assignment outside of section. Ignoring.
Apr 02 02:22:04 ip-10-197-133-18 docker[27826]: Unable to find image 'coreos/subgun' locally
Apr 02 02:22:04 ip-10-197-133-18 docker[27826]: Pulling repository coreos/subgun
core@ip-10-197-133-18 ~ $

and this is the output of fleetctl:

calavera@dcp ~/oss/fleet/foo_Test$ ../bin/fleetctl journal subgun-http.1.service
-- Logs begin at Sun 2014-03-23 21:52:09 UTC, end at Wed 2014-04-02 02:22:29 UTC. --
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:8] Unknown section 'X-Fleet'. Ignoring.
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:9] Assignment outside of section. Ignoring.
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: Starting subgun...
Apr 02 02:22:02 ip-10-197-133-18 systemd[1]: Started subgun.
Apr 02 02:22:03 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:8] Unknown section 'X-Fleet'. Ignoring.
Apr 02 02:22:03 ip-10-197-133-18 systemd[1]: [/run/systemd/system/subgun-http.1.service:9] Assignment outside of section. Ignoring.
Apr 02 02:22:04 ip-10-197-133-18 docker[27826]: Unable to find image 'coreos/subgun' locally
Apr 02 02:22:04 ip-10-197-133-18 docker[27826]: Pulling repository coreos/subgun

[calavera]

Definitely not blocking this issue, but I think we should put all the logic to run commands together somewhere. Right now there are three places where it's duplicated:

1. fleetctl/journal.go
2. fleetctl/ssh.go
3. fleetctl/status.go

[bcwaldon]

@calavera Yes, this patch is not functioning properly. I'm going to merge it and try to refactor the code to your point above ^. I'll add functional testing around these features, too.