cassandra doesn't start while upgrading from 3.0 to 3.2 because nodetool status doesn't work
Affects | Status | Importance | Assigned to | Milestone | ||
---|---|---|---|---|---|---|
Juniper Openstack | Status tracked in Trunk | |||||
R3.0 |
Fix Committed
|
High
|
Unassigned | |||
R3.0.3.x |
Fix Committed
|
High
|
Unassigned | |||
R3.1 |
Fix Committed
|
High
|
Unassigned | |||
R3.2 |
Fix Committed
|
High
|
Unassigned | |||
Trunk |
Fix Committed
|
High
|
Unassigned |
Bug Description
I was upgrading from 3.0.3.1-7~liberty to 3.2.0.0-7~mitaka and I noticed that during the upgrade when fab tries to start cassandra it doesn't work because it's trying to get the nodetool status result and nodetool just responds with : "Cassandra 2.0 and later require Java 7u25 or later", and the fab upgrade would loop like this forever :
2016-12-02 11:22:29:572656: [root@10.87.65.20] out: Configuration file '/etc/cassandra
2016-12-02 11:22:29:572852: [root@10.87.65.20] out: ==> Modified (by you or by a script) since installation.
2016-12-02 11:22:29:573010: [root@10.87.65.20] out: ==> Package distributor has shipped an updated version.
2016-12-02 11:22:29:573088: [root@10.87.65.20] out: ==> Using new file as you requested.
2016-12-02 11:22:29:573158: [root@10.87.65.20] out: Installing new version of config file /etc/cassandra/
2016-12-02 11:22:29:573229: [root@10.87.65.20] out: Installing new version of config file /etc/cassandra/
2016-12-02 11:22:29:577004: [root@10.87.65.20] out: Installing new version of config file /etc/cassandra/
2016-12-02 11:22:29:584668: [root@10.87.65.20] out:
2016-12-02 11:22:29:588236: [root@10.87.65.20] out: Configuration file '/etc/cassandra
2016-12-02 11:22:29:588337: [root@10.87.65.20] out: ==> Modified (by you or by a script) since installation.
2016-12-02 11:22:29:588384: [root@10.87.65.20] out: ==> Package distributor has shipped an updated version.
2016-12-02 11:22:29:588424: [root@10.87.65.20] out: ==> Using new file as you requested.
2016-12-02 11:22:29:588460: [root@10.87.65.20] out: Installing new version of config file /etc/cassandra/
2016-12-02 11:22:29:588497: [root@10.87.65.20] out: vm.max_map_count = 1048575
2016-12-02 11:22:29:652474: [root@10.87.65.20] out: net.ipv4.
2016-12-02 11:22:29:652588: [root@10.87.65.20] out: Processing triggers for python-support (1.0.15) ...
2016-12-02 11:22:29:716448: [root@10.87.65.20] out: Processing triggers for ureadahead (0.100.0-16) ...
2016-12-02 11:22:29:880561: [root@10.87.65.20] out: [localhost] local: service cassandra stop
2016-12-02 11:22:29:912477: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:22:29:920155: [root@10.87.65.20] out: [localhost] local: service cassandra start;sleep 5
2016-12-02 11:22:35:028712: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:22:40:369170: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:22:45:677939: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:22:50:950122: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:22:56:259409: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:23:01:531545: [root@10.87.65.20] out: [localhost] local: sleep 5
2016-12-02 11:23:06:804483: [root@10.87.65.20] out: [localhost] local: sleep 5
I edited the cassandra-env.sh file as recommended here https:/
-if [ "$JVM_VERSION" \< "1.8" ] && [ "$JVM_PATCH_
+if [ "$JVM_VERSION" \< "1.8" ] && [ "$JVM_PATCH_
Then it passes the test and proceed with the upgrade just fine.
It looks like a Cassandra bug and we might need to use the most recent cassandra version to integrate that fix.
I attached the fab upgrade contrail logs and the cassandra system logs for more information.
Guilhem
tags: | added: analytics |
information type: | Proprietary → Public |
tags: | added: csg |
Review in progress for https:/ /review. opencontrail. org/26972
Submitter: Raj Reddy (<email address hidden>)