Return-Path: X-Original-To: apmail-hbase-issues-archive@www.apache.org Delivered-To: apmail-hbase-issues-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 4D64C113BB for ; Mon, 22 Sep 2014 16:54:35 +0000 (UTC) Received: (qmail 39763 invoked by uid 500); 22 Sep 2014 16:54:35 -0000 Delivered-To: apmail-hbase-issues-archive@hbase.apache.org Received: (qmail 39712 invoked by uid 500); 22 Sep 2014 16:54:35 -0000 Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Delivered-To: mailing list issues@hbase.apache.org Received: (qmail 39700 invoked by uid 99); 22 Sep 2014 16:54:35 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 22 Sep 2014 16:54:35 +0000 Date: Mon, 22 Sep 2014 16:54:34 +0000 (UTC) From: "Andrey Stepachev (JIRA)" To: issues@hbase.apache.org Message-ID: In-Reply-To: References: Subject: [jira] [Commented] (HBASE-7767) Get rid of ZKTable, and table enable/disable state in ZK MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HBASE-7767?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14143407#comment-14143407 ] Andrey Stepachev commented on HBASE-7767: ----------------------------------------- [~stack] bq. ++On chicken-and-egg problem, as I see it system tables cannot be disabled so this issue goes away but from code point of view there will be may _if table == 'tables'_ ugly code. bq. ++ Agree that a table for table state is heavyweight (WAL, assignment issues). A table for schema and state would be more useful but it'd be more important than meta even and would have to be assigned before all else... we'd need to move the .tableinfo out of fs and up in here., etc. and table doesn't solve scaling well, it still will be one rs/master which will response for requests (replicas can be here, but they will not be consistent). bq. Over in HBASE-11974,Andrey Stepachev, you talk of proxy? You saying RS should proxy M requests or are you saying RS could learn of table state from master and proxy this to the client throwing TableDisabledException instead of NSRE? (This latter seems interesting... Master could pass table state and other cluster info the back of the heartbeat). that can be made lazy, region server can hold short-lived cache for table info (state, schema) and throttle client requests. And as I proposed in original ticket, RS should throw TDE, not NSRE. > Get rid of ZKTable, and table enable/disable state in ZK > --------------------------------------------------------- > > Key: HBASE-7767 > URL: https://issues.apache.org/jira/browse/HBASE-7767 > Project: HBase > Issue Type: Sub-task > Components: Zookeeper > Affects Versions: 2.0.0 > Reporter: Enis Soztutar > Assignee: Andrey Stepachev > Fix For: 2.0.0 > > Attachments: 0001-HBASE-7767-Get-rid-of-ZKTable-and-table-enable-disab.patch, 0001-rebase.patch, 7767v2.txt, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch, HBASE-7767.patch > > > As discussed table state in zookeeper for enable/disable state breaks our zookeeper contract. It is also very intrusive, used from the client side, master and region servers. We should get rid of it. -- This message was sent by Atlassian JIRA (v6.3.4#6332)