cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aleksey Yeschenko (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-6463) cleanup causes permission problems
Date Tue, 07 Jan 2014 23:15:50 GMT


Aleksey Yeschenko commented on CASSANDRA-6463:

Can't say, b/c I don't know what caused the issue in the first place. A topology change? Increasing
RF on system_auth without running repair afterwards?

Either way, just run repair on every node (just for the system_auth keyspace) - it won't take
much time.

> cleanup causes permission problems
> ----------------------------------
>                 Key: CASSANDRA-6463
>                 URL:
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core, Tools
>         Environment: Windows 7 / Java
>            Reporter: Andreas Schnitzerling
>              Labels: authentication, cleanup, consistency, cql3, newbie, nodetool, permissions
>             Fix For: 2.0.5
> After a cleanup I loose permissions (f.e. SELECT, MODIFY). When I listed the system_auth/permissions
CF after cleanup I recognized, around the half of all permissions lost - BUT: If I list the
permissions table with consistency ALL, all entries appear again and my programm continues
working. Only if I manual trigger that read-repair! That tells me indirect, that system_auth
is reading using consistency ONE, what causes problems after cleanup (?). A good approach
(?) could be to re-read system_auth after permission fail with consistency > ONE to trigger
read-repair once and keep speed.

This message was sent by Atlassian JIRA

View raw message