cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sam Tunnicliffe (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-11942) Cannot process role related query just after restart
Date Thu, 02 Jun 2016 09:07:59 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-11942?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15311996#comment-15311996
] 

Sam Tunnicliffe commented on CASSANDRA-11942:
---------------------------------------------

Roles were introduced in 2.2 and the default role manager requires that all nodes in the cluster
are running at least that version to be sure that the underlying tables are available. The
version a peer is running is learned when messages are exchanged, so the role manager's setup
includes a delay to allow for some exchanges to take place. By default this is 10s and during
that period, any DCL statements, e.g. {{(CREATE|ALTER|GRANT|REVOKE) ROLE...)}}, will be rejected
in the way you described. 

If you're in an environment where you're sure the whole cluster is on the right version and
adding a delay or retry in the client is not an option, during automated testing for instance,
you can reduce the delay period using the system property {{cassandra.superuser_setup_delay_ms}}.
Adding a value for that in {{conf/jvm.options}} or on the command line when starting Cassandra
will override the default value.


> Cannot process role related query just after restart
> ----------------------------------------------------
>
>                 Key: CASSANDRA-11942
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-11942
>             Project: Cassandra
>          Issue Type: Bug
>         Environment: Ubuntu 14.04.4
> Cassandra 3.0.6 (single node)
> Python (2.7) connector with Native protocol v3
>            Reporter: Petr Malik
>
> I get the following error from Python client when executing ALTER USER 'foo' WITH PASSWORD
%s; just after service restart.
> It works if I wait for some 5s before executing the statement.
> From system.log:
> 2016-06-01 22:07:01.458 InvalidRequest: Error from server: code=2200 [Invalid query]
message="Cannot process role related query as the role manager isn't yet setup. This is
>  likely because some of nodes in the cluster are on version 2.1 or earlier. You need
to upgrade all nodes to Cassandra 2.2 or more to use roles."
> INFO  [main] 2016-06-01 22:06:51,637 Server.java:162 - Starting listening for CQL clients
on /127.0.0.1:9042 (unencrypted)...
> WARN  [main] 2016-06-01 22:06:54,646 Slf4JLogger.java:136 - Failed to generate a seed
from SecureRandom within 3 seconds. Not enough entrophy?
> INFO  [main] 2016-06-01 22:06:54,680 CassandraDaemon.java:471 - Not starting RPC server
as requested. Use JMX (StorageService->startRPCServer()) or nodetool (enablethrift) to
start it



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message