Install Couchbase 4 RC over 4 Beta on Ubuntu 14.04 LTS, No Admin console

40-rc
#1

I installed right over Beta by stopping Couchbase and upgrading:

sudo service couchbase-server stop
sudo dpkg -i …

The couchbase Console did not come up, and restarting the service a second time did nothing, but rebooting the entire VM did NOT cause the couchbase console to return. (I thought it was up, but it was just a cached page in my local browser.)

When it was in a bad state, there was no evidence of activity in …/var/lib/couchbase/logs/http_access.log on today’s date, and netstat -a did not list anything listening on port 8091.

error.log had the following entry repeating on a 3 second interval:

[ns_server:error,2015-09-03T09:21:28.375-04:00,ns_1@couchbase1.ramsoft.biz:ns_log<0.201.0>:ns_log:handle_cast:209]unable to notify listeners because of badarg

info.log contains a strange crash:

[ns_server:info,2015-09-03T09:22:22.158-04:00,ns_1@couchbase1.ramsoft.biz:ns_server_sup<0.194.0>:dir_size:start_link:39]Starting quick version of dir_size with program name: godu
[user:info,2015-09-03T09:22:22.171-04:00,ns_1@couchbase1.ramsoft.biz:<0.200.0>:ns_log:crash_consumption_loop:70]Service ‘ns_server’ exited with status 1. Restarting. Messages: [os_mon] cpu supervisor port (cpu_sup): Erlang has closed
{“Kernel pid terminated”,application_controller,"{application_start_failure,ns_server,{{shutdown,{failed_to_start_child,ns_server_nodes_sup,{shutdown,{failed_to_start_child,ns_server_sup,{shutdown,{failed_to_start_child,index_settings_manager,{badarg,[{dict,fetch,[<<“indexer.settings.log_level”>>,{dict,7,16,16,8,80,48,{[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[]},{{[],[],[[<<“indexer.settings.compaction.interval”>>|<<“00:00,00:00”>>]],[[<<“indexer.settings.persisted_snapshot.interval”>>|30000]],[[<<“indexer.settings.compaction.min_frag”>>|30]],[[<<“indexer.settings.inmemory_snapshot.interval”>>|200]],[],[[<<“indexer.settings.max_cpu_percent”>>|400]],[],[],[[<<“indexer.settings.recovery.max_rollbacks”>>|5]],[[<<“indexer.settings.memory_quota”>>|536870912]],[],[],[],[]}}}],[{file,“dict.erl”},{line,126}]},{index_settings_manager,’-lens_get_many/2-lc$^0/1-0-’,2,[{file,“src/index_settings_manager.erl”},{line,289}]},{index_settings_manager,’-lens_get_many/2-lc$^0/1-0-’,2,[{file,“src/index_settings_manager.erl”},{line,289}]},{index_settings_manager,’-lens_get_many/2-lc$^0/1-0-’,2,[{file,“src/index_settings_manager.erl”},{line,289}]},{index_settings_manager,’-lens_get_many/2-lc$^0/1-0-’,2,[{file,“src/index_settings_manager.erl”},{line,289}]},{index_settings_manager,do_populate_ets_table,1,[{file,“src/index_settings_manager.erl”},{line,168}]},{work_queue,init,1,[{file,“src/work_queue.erl”},{line,52}]},{gen_server,init_it,6,[{file,“gen_server.erl”},{line,304}]}]}}}}}}},{ns_server,start,[normal,[]]}}}"}

Crash dump was written to: erl_crash.dump.1441286218.21327.ns_server
Kernel pid terminated (application_controller) ({application_start_failure,ns_server,{{shutdown,{failed_to_start_child,ns_server_nodes_sup,{shutdown,{failed_to_start_child,ns_server_sup,{shutdown,{f
[ns_server:error,2015-09-03T09:22:22.172-04:00,ns_1@couchbase1.ramsoft.biz:ns_log<0.199.0>:ns_log:handle_cast:209]unable to notify listeners because of badarg
[ns_server:warn,2015-09-03T09:22:22.179-04:00,ns_1@couchbase1.ramsoft.biz:ns_config_isasl_sync<0.201.0>:ns_memcached:connect:1283]Unable to connect: {error,{badmatch,{error,econnrefused}}}, retrying.
[ns_server:info,2015-09-03T09:22:22.354-04:00,ns_1@couchbase1.ramsoft.biz:ns_couchdb_port<0.178.0>:ns_port_server:log:210]ns_couchdb<0.178.0>: 25672: Booted. Waiting for shutdown request
ns_couchdb<0.178.0>: working as port

[user:info,2015-09-03T09:22:23.189-04:00,ns_1@couchbase1.ramsoft.biz:ns_cookie_manager<0.141.0>:ns_cookie_manager:do_cookie_sync:130]Node ‘ns_1@couchbase1.ramsoft.biz’ synchronized otp cookie xxawafcgveyjueul from cluster
[user:info,2015-09-03T09:22:23.286-04:00,ns_1@couchbase1.ramsoft.biz:mb_master<0.252.0>:mb_master:init:86]I’m the only node, so I’m the master.
[ns_server:info,2015-09-03T09:22:23.339-04:00,ns_1@couchbase1.ramsoft.biz:mb_master<0.252.0>:mb_master:terminate:299]Synchronously shutting down child mb_master_sup
[ns_server:info,2015-09-03T09:22:23.348-04:00,ns_1@couchbase1.ramsoft.biz:ns_couchdb_port<0.178.0>:ns_port_server:log:210]ns_couchdb<0.178.0>: 25672: got shutdown request. Exiting
ns_couchdb<0.178.0>: [os_mon] memory supervisor port (memsup): Erlang has closed
ns_couchdb<0.178.0>: [os_mon] cpu supervisor port (cpu_sup): Erlang has closed

#2

Couchbase needs to listen on more ports than just 8091, so you should check the others listed in the documentation.

Having said that, have you tried uninstalling the Beta 4 package, and then installing RC0 cleanly?

#3

Probably would be a good idea, but dpkg -i supports an inplace upgrade, so if there’s an inherent issue in doing so, it should warn people.

logs on onedrive (76mb, .zip): http://1drv.ms/1ik4M92

#4

I may be wrong, but given that Beta4 isn’t an official release I doubt that upgrading from it is supported.

#5

I think most developers like to know if their stuff is broken, and to have the logs. That’s why I bring it up. Also other people might try it and find something else, that is different than I found.

Rather than staying stuck if anyone else hits this issue, I suggest people do a full uninstall, zip or tar up the old logs in case they are useful to the developers, and then remove /opt/couchbase completely, to ensure a clean 4.0 RC0 experience.

W