accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-2918) replication makes tservers quit
Date Tue, 17 Jun 2014 15:33:03 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-2918?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14033927#comment-14033927
] 

Josh Elser commented on ACCUMULO-2918:
--------------------------------------

Ok, so:

# Do I need to do something that automatically creates these nodes in the tserver to account
for the churn within 1.7.0-SNAPSHOT? My inclination is no (which is why you ran into this
:) ). It just adds a little complexity to the tserver which, after we release the code, is
useless.
# I need to find upgrade code inside the master (assuming that's where it lives). Creating
this node in the tserver would also satisfy this case, but then we're also baking in upgrade
logic into the server process which is yuck.

Thoughts?

> replication makes tservers quit
> -------------------------------
>
>                 Key: ACCUMULO-2918
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2918
>             Project: Accumulo
>          Issue Type: Bug
>          Components: replication
>         Environment: starting up 1.7.0 after replication was commited
>            Reporter: Eric Newton
>            Assignee: Josh Elser
>
> replication needs some upgrade code to make some nodes in zookeeper
> Here's the error message:
> {noformat}
> KeeperErrorCode = NoNode for /accumulo/2f3173c3-c02f-4e3a-b01c-7f721d4a9f00/replication/tservers/host:9997
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message