These are chat archives for dereneaton/ipyrad

25th
Jul 2017
Wind-ant
@Wind-ant
Jul 25 2017 00:13
thank you @isaacovercast
joqb
@joqb
Jul 25 2017 14:34

Hi @isaacovercast, @dereneaton,
Our system got updated now after a complete reinstall of conda and ipyrad I'm getting this error:

Hub connection request timed out

Any idea?

Isaac Overcast
@isaacovercast
Jul 25 2017 14:42
Have you tried starting ipcluster by hand?
ipcluster start -n 4
This error is an ipyparallel timeout error, it's having trouble starting the cluster for some reason.
joqb
@joqb
Jul 25 2017 15:31

Yes I had tried this

ipcluster start --n 20 --daemonize

But now I tried what you suggested, here is how it reacted:

bash-4.2$ ipcluster start -n 4
2017-07-25 17:29:09.165 [IPClusterStart] Starting ipcluster with [daemon=False]
2017-07-25 17:29:09.168 [IPClusterStart] Creating pid file: /home/brassac/.ipython/profile_default/pid/ipcluster.pid
2017-07-25 17:29:09.170 [IPClusterStart] Starting Controller with LocalControllerLauncher
2017-07-25 17:29:10.182 [IPClusterStart] Starting 4 Engines with LocalEngineSetLauncher
2017-07-25 17:29:13.798 [IPClusterStart] ERROR | IPython cluster: stopping
2017-07-25 17:29:13.799 [IPClusterStart] Stopping Engines...
2017-07-25 17:29:16.803 [IPClusterStart] Removing pid file: /home/brassac/.ipython/profile_default/pid/ipcluster.pid
Deren Eaton
@dereneaton
Jul 25 2017 15:38
Looks like it started very quickly. If you add the argument '--ipcluster' to your ipyrad command it will use your daemonized cluster to run the job.
If you are trying to connect to cores across multiple nodes then you need a few more arguments.
The hub timeout may just be a one time problem caused by the system update. If you try again it may work fine.
joqb
@joqb
Jul 25 2017 16:04
I tried with the --ipcluster argument but it's the same: hub timeout.
Is it normal that the engines are being stopped so quickly?
Deren Eaton
@dereneaton
Jul 25 2017 16:06
Oh, I assumed you stopped it.
joqb
@joqb
Jul 25 2017 17:17
No I didn't, it's getting the error after a couple of seconds and then dies. However something still runs in the background
 358 brassac   20   0  274964  32628   1552 S  12.3  0.0   8:25.34 ipcluster
32225 brassac   20   0  274952  32612   1552 S  11.3  0.0   7:38.09 ipcluster
 1275 brassac   20   0  274952  32616   1552 R  10.3  0.0   7:20.52 ipcluster
  382 brassac   20   0  589136  40624   6404 S   0.3  0.0   0:02.60 python
 1276 brassac   20   0  452348  53284  10788 S   0.3  0.0   0:06.30 python
 1280 brassac   20   0  589136  40628   6404 S   0.3  0.0   0:02.39 python
 1323 brassac   20   0  589140  40636   6404 S   0.3  0.0   0:02.66 python
 1395 brassac   20   0  589140  38592   6404 S   0.3  0.0   0:02.67 python
 1418 brassac   20   0  589136  40628   6404 S   0.3  0.0   0:02.50 python
 1427 brassac   20   0  589136  38588   6404 S   0.3  0.0   0:02.52 python
Deren Eaton
@dereneaton
Jul 25 2017 17:19
Run ipcluster stop to stop it.
It may be stopping bc there is already one running on the default profile.
joqb
@joqb
Jul 25 2017 17:21
Yeah I know but this doesn't seem to do anything
-bash-4.2$ ipcluster stop
2017-07-25 19:20:16.672 [IPClusterStop] CRITICAL | Could not read pid file, cluster is probably not running.
Deren Eaton
@dereneaton
Jul 25 2017 17:21
killall ipcluster
Should do the trick
joqb
@joqb
Jul 25 2017 17:26
It did the trick and I killed all but ipcluster start -n 4 still stops after a couple of seconds
Deren Eaton
@dereneaton
Jul 25 2017 17:29
Is it possible your hard disk is full?
I'm just on my phone right now, so I can't check, but try ipcluster --help and see if there is an option to turn on more debugging to get a better explanation of the error.
joqb
@joqb
Jul 25 2017 18:22
Thanks for the tips, I hope that by tomorrow my ITs will have reacted and maybe have a solution for me...