Open davidsteinsland opened 9 years ago
Are you using capistrano or similar tools for deployment ?
Yes, I'm using DeployBot.
Does deploybot have a shared folder ? Pid files should reside in a shared folder, so that file persists between deploy.
On Monday, 9 November 2015, David Steinsland notifications@github.com wrote:
Yes, I'm using DeployBot.
— Reply to this email directly or view it on GitHub https://github.com/wa0x6e/Cake-Resque/issues/96#issuecomment-155051177.
Yes, all that is taken care off.
The tmp folder is shared between deployment ?
On Tuesday, 10 November 2015, David Steinsland notifications@github.com wrote:
Yes, all that is taken care off.
— Reply to this email directly or view it on GitHub https://github.com/wa0x6e/Cake-Resque/issues/96#issuecomment-155150131.
Again, yes. Permissions are OK. Problem is not with my setup. As I said initially, the processes are started with the correct environment variables. So any failure should be logged, and it isnt.
Ca you post the full deployment log for DeployBot ?
@davidsteinsland
I've specified an alternate location of the pid files, and they are showing up in the commands:
You didn't mentioned it specifically, but if you mean CakeResque.Resque.tmpdir
: I did a analysis recently and found out that these PID files are only very short lived and used to check if the worker properly started and are then removed from the file system:
_checkStartedWorker
_checkStartedWorker
: pidfile removedOnly within redis the PIDs are permanently persisted for the lifetime of a worker.
I've specified an alternate location of the pid files, and they are showing up in the commands:
No file with that path exists, nor does the logs contain any information as to why it doesn't. When no pid file exists, I cannot stop the workers via the console shell, I have to kill them manually.
On a second note, the
LOGHANDLERTARGET
is also set:This file doesn't exist either.