[ https://issues.apache.org/jira/browse/HBASE-22012?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16814077#comment-16814077 ]
Shardul Singh edited comment on HBASE-22012 at 4/10/19 6:02 AM: ---------------------------------------------------------------- [~elserj] Hi josh, I think just adding failsafe when we're checking for a quota that has the DisableTable violation policy don't filter in *tablesToRemove* won't work because after retention period for region report expires it removes the entry from the cache regionSizes, but since region report size are not present in cache, *targetSnapshot* in quotaObserverChore will give 0B as usage and table still get's enabled during *processTablesWithQuotas* step. I think it's better to not remove the entries from cache regionSizes if table is in *violation* and has *DISABLE* policy set on it. was (Author: shardulsingh): [~elserj] Hi josh, I think just adding failsafe when we're checking for a quota that has the DisableTable violation policy don't filter in *tablesToRemove* won't work because after retention period for region report expires it removes the entry from the cache regionSizes, but since region report size are not present in cache, *targetSnapshot* in quotaObserverChore will give 0B as usage and table still get's enabled during *processTablesWithQuotas* step*.* I think it's better to not remove the entries from cache regionSizes if table is in *violation* and has *DISABLE* policy set on it. > SpaceQuota DisableTableViolationPolicy will cause cycles of enable/disable > table > -------------------------------------------------------------------------------- > > Key: HBASE-22012 > URL: https://issues.apache.org/jira/browse/HBASE-22012 > Project: HBase > Issue Type: Bug > Affects Versions: 3.0.0, 2.2.1 > Reporter: Ajeet Rai > Assignee: Nihal Jain > Priority: Major > > Space Quota: Policy state is getting changed from disable to Observance after > sometime automatically. > Steps: > 1: Create a table with space quota policy as Disable > 2: Put some data so that table state is in space quota violation > 3: So observe that table state is in violation > 4: Now wait for some time > 5: Observe that after some time table state is changing to to Observance > however table is still disabled > edit (elserj): The table is automatically moved back from the violation state > because of the code added that tried to ride over RITs. When a Region is not > online (whether normally or abnormally), the RegionSizeReports are not sent > from RS to Master. Eventually, enough Regions are not reported which dips > below the acceptable threshold and we automatically move the table back to > the "acceptable" space quota state (not in violation). We could skip this > failsafe when we're checking for a quota that has the DisableTable violation > policy. -- This message was sent by Atlassian JIRA (v7.6.3#76005)