From notifications-return-47014-archive-asf-public=cust-asf.ponee.io@accumulo.apache.org Wed Jan 2 22:19:07 2019 Return-Path: X-Original-To: archive-asf-public@cust-asf.ponee.io Delivered-To: archive-asf-public@cust-asf.ponee.io Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by mx-eu-01.ponee.io (Postfix) with SMTP id 066C918067A for ; Wed, 2 Jan 2019 22:19:06 +0100 (CET) Received: (qmail 43938 invoked by uid 500); 2 Jan 2019 21:19:06 -0000 Mailing-List: contact notifications-help@accumulo.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: jira@apache.org Delivered-To: mailing list notifications@accumulo.apache.org Received: (qmail 43927 invoked by uid 99); 2 Jan 2019 21:19:06 -0000 Received: from ec2-52-202-80-70.compute-1.amazonaws.com (HELO gitbox.apache.org) (52.202.80.70) by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 02 Jan 2019 21:19:06 +0000 From: GitBox To: notifications@accumulo.apache.org Subject: [GitHub] joshelser commented on issue #868: Make tserver start repl services when configured Message-ID: <154646394550.22675.10440360629424688736.gitbox@gitbox.apache.org> Date: Wed, 02 Jan 2019 21:19:05 -0000 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit joshelser commented on issue #868: Make tserver start repl services when configured URL: https://github.com/apache/accumulo/pull/868#issuecomment-450988255 > I am not too familiar with replication but it's properties might be used in tserver and master. That's correct. There's logic across master and tabletserver for replication. The intent to have its own prefix was to keep all replication configuration in "one place". > just have a general message in the docs like zk mutable: yes but requires a restart +1 > With this change, we could throw an illegal state exception if the user sets a property on the fly, expecting data to get replicated. I am not sure where that would be done though. That would be client side validation, right? I suppose it's not unreasonable for this case to let client-side validation be sufficient (e.g. it's not bad if a user circumvents our own code and tries to set a zk-mutable property -- worst case, it's just ineffectual). That sounds like a good idea. Even if not an exception, a WARN would go a long way for operators to realize they straying from the "norm" ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: users@infra.apache.org With regards, Apache Git Services