[ 
https://issues.apache.org/jira/browse/PHOENIX-5034?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16708099#comment-16708099
 ] 

ASF GitHub Bot commented on PHOENIX-5034:
-----------------------------------------

Github user karanmehta93 commented on a diff in the pull request:

    https://github.com/apache/phoenix/pull/409#discussion_r238509787
  
    --- Diff: 
phoenix-core/src/main/java/org/apache/phoenix/log/QueryLogger.java ---
    @@ -92,9 +92,17 @@ public boolean isSynced(){
             }
         };
     
    -    public static QueryLogger getInstance(PhoenixConnection connection, 
boolean isSystemTable) {
    -        if (connection.getLogLevel() == LogLevel.OFF || isSystemTable || 
ThreadLocalRandom.current()
    -                .nextDouble() > connection.getLogSamplingRate()) { return 
NO_OP_INSTANCE; }
    +    public static QueryLogger getInstance(PhoenixConnection connection, 
boolean isSystemTable,
    +        boolean criticalStatement) {
    +        // always log critical statements (DROP,ALTER for now)
    +        // do not log anything when loglevel is off.
    +        // do not log systemTable statement.
    +        // do sampling on other statements based on configured 
perncentage, 1% by default.
    --- End diff --
    
    nit: perncentage


> Log all critical statements in SYSTEM.LOG table.
> ------------------------------------------------
>
>                 Key: PHOENIX-5034
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-5034
>             Project: Phoenix
>          Issue Type: Improvement
>            Reporter: Xu Cang
>            Assignee: Xu Cang
>            Priority: Minor
>         Attachments: PHOENIX-5034-4.x-HBase-1.3.001.patch, 
> PHOENIX-5034-4.x-HBase-1.3.002.patch, PHOENIX-5034-4.x-HBase-1.3.003.patch, 
> PHOENIX-5034-4.x-HBase-1.3.004.patch, PHOENIX-5034-4.x-HBase-1.3.005.patch
>
>
> In production, sometimes engineers see table got dropped unexpectedly. It's 
> not easy to SCAN raw table from HBase itself to understand what happened and 
> when the table get dropped.
> Since we already have SYSTEM.LOG query log facility in Phoenix that sampling 
> query statement (log 1% statement by default). It's good to always log 
> critical statements such as "DROP" or "ALTER" statements.
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to