hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Purtell (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-18054) log when we add/remove failed servers in client
Date Thu, 01 Jun 2017 23:50:04 GMT

    [ https://issues.apache.org/jira/browse/HBASE-18054?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16033925#comment-16033925
] 

Andrew Purtell commented on HBASE-18054:
----------------------------------------

bq. Thanks. Now let's look at the patch.

{code}
51	    if (LOG.isDebugEnabled()) {
52	      LOG.debug("Added failed server with address " + address.toString() + " to list");
53	    }
{code}

This doesn't tell us more than we can deduce by a later log indicating the server is in the
failed server list. We should add why the server has been added to the failed server list.
Perhaps addToFailedServers() should get a new additional argument, a string containing the
reason, that all callers can supply?  

> log when we add/remove failed servers in client
> -----------------------------------------------
>
>                 Key: HBASE-18054
>                 URL: https://issues.apache.org/jira/browse/HBASE-18054
>             Project: HBase
>          Issue Type: Bug
>          Components: Client, Operability
>    Affects Versions: 1.3.0
>            Reporter: Sean Busbey
>            Assignee: Ali
>         Attachments: HBASE-18054.patch, HBASE-18054.v2.master.patch, HBASE-18054.v3.master.patch
>
>
> Currently we log if a server is in the failed server list when we go to connect to it,
but we don't log anything about when the server got into the list.
> This means we have to search the log for errors involving the same server name that (hopefully)
managed to get into the log within {{FAILED_SERVER_EXPIRY_KEY}} milliseconds earlier (default
2 seconds).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message