[
https://issues.apache.org/jira/browse/MESOS-2340?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14652241#comment-14652241
]
Marco Massenzio commented on MESOS-2340:
----------------------------------------
Please don't use this ticket to file errors - it's much better to ask for help
in the dev@ mailing list.
Having said that, what you are attaching here is most definitely *not* the
error cause: this log in fact, confirms that the Master was able to
successfully decode the MasterInfo in Zookeeper (you probably still have a 0.23
Master running?).
What error do you see? what is the symptom? and you may want to upload the full
log output (ideally, as an attachment).
Finally, if you really have installed "mesosphere" (did you mean DCOS?) and not
Apache Mesos, you may want to file a request support there instead.
FYI - this is what a 0.24 Master emits when starting up:
{noformat}
$ ./bin/mesos-master.sh --zk=zk://localhost:2181/test/report --quorum=1
--work_dir=/tmp/report
...
I0803 11:33:41.037178 222056448 group.cpp:331] Group process
(group(2)@10.0.77.243:5050) connected to ZooKeeper
I0803 11:33:41.037196 220446720 group.cpp:331] Group process
(group(1)@10.0.77.243:5050) connected to ZooKeeper
I0803 11:33:41.037204 222056448 group.cpp:805] Syncing group operations: queue
size (joins, cancels, datas) = (1, 0, 0)
I0803 11:33:41.037217 220446720 group.cpp:805] Syncing group operations: queue
size (joins, cancels, datas) = (0, 0, 0)
I0803 11:33:41.037250 220446720 group.cpp:403] Trying to create path
'/test/report/log_replicas' in ZooKeeper
I0803 11:33:41.037250 222056448 group.cpp:403] Trying to create path
'/test/report/log_replicas' in ZooKeeper
2015-08-03 11:33:41,037:3806(0x10d8f3000):ZOO_INFO@check_events@1750: session
establishment complete on server [::1:2181], sessionId=0x14ed7154df50023,
negotiated timeout=10000
I0803 11:33:41.038208 220983296 group.cpp:331] Group process
(group(3)@10.0.77.243:5050) connected to ZooKeeper
I0803 11:33:41.038229 220983296 group.cpp:805] Syncing group operations: queue
size (joins, cancels, datas) = (0, 0, 0)
I0803 11:33:41.038240 220983296 group.cpp:403] Trying to create path
'/test/report' in ZooKeeper
2015-08-03 11:33:41,038:3806(0x10d976000):ZOO_INFO@check_events@1750: session
establishment complete on server [::1:2181], sessionId=0x14ed7154df50024,
negotiated timeout=10000
I0803 11:33:41.038566 223666176 group.cpp:331] Group process
(group(4)@10.0.77.243:5050) connected to ZooKeeper
I0803 11:33:41.038595 223666176 group.cpp:805] Syncing group operations: queue
size (joins, cancels, datas) = (0, 0, 0)
I0803 11:33:41.038614 223666176 group.cpp:403] Trying to create path
'/test/report' in ZooKeeper
I0803 11:33:41.058315 222593024 contender.cpp:149] Joining the ZK group
I0803 11:33:41.078806 222593024 contender.cpp:265] New candidate (id='1') has
entered the contest for leadership
I0803 11:33:41.079265 221519872 detector.cpp:156] Detected a new leader:
(id='1')
I0803 11:33:41.079429 223129600 group.cpp:674] Trying to get
'/test/report/json.info_0000000001' in ZooKeeper
I0803 11:33:41.079447 219910144 network.hpp:415] ZooKeeper group memberships
changed
I0803 11:33:41.079527 222593024 group.cpp:674] Trying to get
'/test/report/log_replicas/0000000000' in ZooKeeper
I0803 11:33:41.081306 220983296 detector.cpp:481] A new leading master
([email protected]:5050) is detected
I0803 11:33:41.081464 219910144 network.hpp:463] ZooKeeper group PIDs: {
log-replica(1)@10.0.77.243:5050 }
I0803 11:33:41.081482 223129600 master.cpp:1495] The newly elected leader is
[email protected]:5050 with id 20150803-113340-4081909770-5050-3806
I0803 11:33:41.082784 223129600 master.cpp:1508] Elected as the leading master!
I0803 11:33:41.083602 223129600 master.cpp:1278] Recovering from registrar
I0803 11:33:41.085345 222056448 registrar.cpp:313] Recovering registrar
{noformat}
the next Master detects it:
{noformat}
$ ./bin/mesos-master.sh --zk=zk://localhost:2181/test/report --quorum=1
--work_dir=/tmp/report2 --port=5051
...
I0803 11:37:27.356122 317771776 group.cpp:331] Group process
(group(4)@10.0.77.243:5051) connected to ZooKeeper
I0803 11:37:27.356145 317771776 group.cpp:805] Syncing group operations: queue
size (joins, cancels, datas) = (0, 0, 0)
I0803 11:37:27.356158 317771776 group.cpp:403] Trying to create path
'/test/report' in ZooKeeper
I0803 11:37:27.360528 316162048 network.hpp:415] ZooKeeper group memberships
changed
I0803 11:37:27.360651 315088896 group.cpp:674] Trying to get
'/test/report/log_replicas/0000000000' in ZooKeeper
I0803 11:37:27.360689 317771776 detector.cpp:156] Detected a new leader:
(id='1')
I0803 11:37:27.360949 317771776 group.cpp:674] Trying to get
'/test/report/json.info_0000000001' in ZooKeeper
I0803 11:37:27.361369 315088896 group.cpp:674] Trying to get
'/test/report/log_replicas/0000000001' in ZooKeeper
I0803 11:37:27.362244 317235200 network.hpp:463] ZooKeeper group PIDs: {
log-replica(1)@10.0.77.243:5050, log-replica(1)@10.0.77.243:5051 }
I0803 11:37:27.363267 318308352 master.cpp:377] Master
20150803-113727-4081909770-5051-3865 (10.0.77.243) started on 10.0.77.243:5051
...
{noformat}
if you then run a 0.23 Mesos Master, it all "just works":
{noformat}
$ cd ../build-23
$ ./bin/mesos-master.sh --zk=zk://localhost:2181/test/report --quorum=1
--work_dir=/tmp/report3 --port=5052
I0803 11:39:19.031148 2039251712 main.cpp:181] Build: 2015-07-23 14:21:55 by
marco
I0803 11:39:19.031390 2039251712 main.cpp:183] Version: 0.23.0
I0803 11:39:19.031399 2039251712 main.cpp:186] Git tag: 0.23.0
I0803 11:39:19.031406 2039251712 main.cpp:190] Git SHA:
4ce5475346a0abb7ef4b7ffc9836c5836d7c7a66
I0803 11:39:19.031702 2039251712 main.cpp:204] Using 'HierarchicalDRF' allocator
...
I0803 11:39:19.110242 292507648 detector.cpp:481] A new leading master
([email protected]:5050) is detected
I0803 11:39:19.128594 291434496 contender.cpp:131] Joining the ZK group
I0803 11:39:19.128746 289288192 master.cpp:1481] The newly elected leader is
[email protected]:5050 with id 20150803-113340-4081909770-5050-3806
I0803 11:39:19.129890 291434496 contender.cpp:247] New candidate (id='3') has
entered the contest for leadership
...
{noformat}
Please double check your logs if you can find an error message.
> Add ability to decode JSON serialized MasterInfo from ZK
> --------------------------------------------------------
>
> Key: MESOS-2340
> URL: https://issues.apache.org/jira/browse/MESOS-2340
> Project: Mesos
> Issue Type: Improvement
> Components: leader election
> Reporter: Zameer Manji
> Assignee: Marco Massenzio
> Labels: mesosphere
> Fix For: 0.23.0
>
>
> Currently to discover the master a client needs the ZK node location and
> access to the MasterInfo protobuf so it can deserialize the binary blob in
> the node.
> I think it would be nice to publish JSON (like Twitter's ServerSets) so
> clients are not tied to protobuf to do service discovery.
> This ticket is an intermediate (compatibility) step: we add in {{0.23}} the
> ability for the {{Detector}} to "understand" JSON **alongside** Protobuf
> serialized format; this makes it compatible with both earlier versions, as
> well a future one (most likely, {{0.24}}) that will write the {{MasterInfo}}
> information in JSON format.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
