Can't connect to slave

[Philipp Sessler]

Hi,

I'm having trouble setting up the cluster. Using only the master node as slave works fine but adding another node as slave produces the following error:

node5 discouser 83 ( bin ) $ ./disco debug
Erlang/OTP 17 [erts-6.4] [source] [64-bit] [smp:120:120] [async-threads:10] [hipe] [kernel-poll:false]

*** ERROR: Shell process terminated! (^G to start new job) ***


node5 discouser 86 ( bin ) $ ./disco nodaemon
Erlang/OTP 17 [erts-6.4] [source] [64-bit] [smp:120:120] [async-threads:10] [hipe] [kernel-poll:true]

Eshell V6.4  (abort with ^G)
(disco_8989_master@node5)1> 15:57:41.701 [info] Application lager started on node disco_8989_master@node5
15:57:41.755 [info] Application inets started on node disco_8989_master@node5
15:57:41.758 [info] DISCO BOOTS
15:57:41.760 [info] Disco proxy disabled
15:57:41.763 [info] DDFS master starts
15:57:41.767 [info] Event server starts
15:57:41.769 [info] Disco config starts
15:57:41.772 [info] DISCO SERVER STARTS
15:57:41.774 [info] Fair scheduler starts
15:57:41.774 [info] Scheduler uses fair policy
15:57:41.776 [info] Fair scheduler: Fair policy
15:57:41.780 [info] Config table updated
15:57:41.792 [info] Starting node "disco_8989_slave" on "localhost" ("localhost")
15:57:41.801 [info] web server (mochiweb) starts
15:57:41.801 [info] Application disco started on node disco_8989_master@node5
15:57:42.359 [info] ddfs_node initialized on disco_8989_master@node5 with volumes: ["vol0"]
15:57:42.360 [info] lock_server starts on disco_8989_slave@localhost
15:57:42.362 [info] ddfs_node initialized on disco_8989_slave@localhost with volumes: ["vol0"]
15:57:42.363 [info] ddfs_node starts on disco_8989_master@node5
15:57:42.363 [info] Node started at disco_8989_slave@localhost (reporting as disco_8989_master@node5) on "localhost"
15:57:42.370 [info] Tempgc: error listing "/vol/fob-vol6/mi13/discouser/testprj/tools/disco/root/data/localhost": {error,enoent}
15:57:42.370 [info] Tempgc: one pass completed on disco_8989_slave@localhost
15:57:42.378 [info] Started ddfs_put at disco_8989_slave@localhost on port 8990
15:57:42.379 [info] ddfs_node starts on disco_8989_slave@localhost
15:57:42.435 [info] Node started at disco_8989_slave@localhost (reporting as disco_8989_slave@localhost) on "localhost"
15:58:11.783 [warning] net_adm:names() failed for "node6": {badrpc,timeout}
16:00:41.785 [warning] net_adm:names() failed for "node6": {badrpc,timeout}
16:02:41.774 [info] GC: initializing
16:02:55.778 [error] GC: unable to start: too_many_failed_nodes
16:02:55.778 [info] GC terminated.
16:03:11.787 [warning] net_adm:names() failed for "node6": {badrpc,timeout}

When i try to run disco debug the following error occurs:

node5 discouser 83 ( bin ) $ ./disco debug
Erlang/OTP 17 [erts-6.4] [source] [64-bit] [smp:120:120] [async-threads:10] [hipe] [kernel-poll:false]

*** ERROR: Shell process terminated! (^G to start new job) ***

Whereas ssh localhost erl works just fine:

node5 discouser 92 ( bin ) $ ssh localhost erl
Eshell V6.4  (abort with ^G)
1>

node5 ist the master node and the node6 the slave node. They have a shared home directory (so all keys and the erlang cookie are identical). I can also ssh into each other by just typing ssh node6 / ssh node5 without specifying any password. Also, they are resolved by DNS and not by /etc/hosts.

Any thoughts on this would be much appreciated :)

Thanks!

[Shayan Pooya]

Hi,

Would you please try this and see if it works?
https://github.com/pooya/discostuff/blob/master/healthcheck.sh

> --
> You received this message because you are subscribed to the Google Groups
> "Disco-development" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to disco-dev+...@googlegroups.com.
> To post to this group, send email to disc...@googlegroups.com.
> Visit this group at http://groups.google.com/group/disco-dev.
> For more options, visit https://groups.google.com/d/optout.