[13:00] hi xnox, jodh I need to export an SSL cert before respawning the service..with a pre-start stanza it's not working [13:01] like: pre-start script [13:01] export SSL_CERT=/path [13:01] echo 'text' [13:01] end script [13:03] Mariano9: things you export in pre-start, don't propagate elsewhere. [13:04] Mariano9: you can use the global "export SSL_CERT=/path" upstart stanza, then it will be available during all stages (pre-start, script/exec, post-start, pre-stop, post-stop, etc.) [13:04] ok..let's see [13:05] is "env" stanza what you mean? [13:08] xnox: [13:08] Mariano9: yeah, that. =) see $ man 5 init [13:12] ok, tried with env and export stanzas, none worked for my example [13:13] which user is the one who runs upstart scripts? [14:09] xnox: could you take a look at my redis-commander upstart script? I'm trying to export an SSL cert path to allow twitter sensu hander to pick it up [14:10] https://gist.github.com/Mariano-gon/8916561 [14:14] Mariano9: what's running on port 3030? [14:14] Mariano9: redis-commander itself or something else? [14:14] sensu client [14:15] Mariano9: ok. looks all correct. But the environmet variable only affect redis-commander. [14:16] i think you're right..(thinking here).. [14:16] Mariano9: since nc, doesn't do ssl =) [14:16] Mariano9: http://serverfault.com/questions/102032/connecting-to-https-with-netcat-nc [14:16] Mariano9: if sensu client is expecting an ssl connection, use openssl s_client -connect instead. [14:17] what's the certificate for? and who should be validating requests from where using it? [14:17] the client-server sensu uses twitter.rb (a handler script) to connect with twitter API and send a direct message to X account [14:17] this connection to the API needs SSL certification [14:18] Mariano9: thus the certificate should be in the environment of the sensu process, and has nothing to do with the above job. [14:19] yes, now I realize that [14:19] :) [14:19] sorry [14:19] =))) [14:19] other one [14:19] any ideas about rackspace plugin script? [14:20] that one does have an env issue [14:20] re:bundler? didn't play with that yet. [14:20] yeap [14:20] ok [14:20] let me know if there's something I can help you with [14:58] I have an upstart job stuck in "stop/killed" state (ureadhead-other), with init (pid 1) still holding an open log-file for the job, which I need closing (the log-file is inside a chroot that needs umount-ing). The job process no longer exists. How do I proceed? === mc_bluebeard_ is now known as mc_bluebeard === mc_bluebeard_ is now known as mc_bluebeard [23:09] Ubuntu and Canonical are missing the opportunity for upstart in Debian, much thanks to the CLA [23:09] The CLA is harming adoption of Upstart [23:10] Debian could have went with Upstart, but now they are going with systemd instead