accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-3842) [UMBRELLA] Remove non-transient data from ZooKeeper
Date Sat, 23 May 2015 19:19:17 GMT


Josh Elser commented on ACCUMULO-3842:

Caught up one some procv2 HBase stuff

HBASE-13571 deals with schema updates. This is done by re-opening every region. That isn't
relevant for what we're talking about here.

HBASE-13687 and HBASE-13688 mention that there is a "missing piece" that might be relevant,
but the information is lacking.

On the original design docs, I see

Multi-Machine Procedures and Timeouts
Operations like Snapshots or ACLs cache updates requires a bit of coordination across multiple
machine. To do that the procedure will send a message (may be done as poll via heartbeat)
to each machine required by the procedure and will wait until each one respond. The procedure
can have a timeout that will trigger a failure of the procedure causing the rollback.

This doesn't seem like anything novel that trying to adopt procv2 would gain us that we couldn't
already do with FATE. I'm happy to entertain a conversation if I missed something, but, from
what I've read so far, I don't see a reason why we'd want to adopt procv2 presently.

> [UMBRELLA] Remove non-transient data from ZooKeeper
> ---------------------------------------------------
>                 Key: ACCUMULO-3842
>                 URL:
>             Project: Accumulo
>          Issue Type: Improvement
>          Components: client, tserver
>            Reporter: Josh Elser
>             Fix For: 1.8.0
> Wanted to start brainstorming about this.
> We store a lot of persistent data in ZooKeeper that would better stored in something
backed by HDFS. ZooKeeper can be a very convenient place to store persisted data so that it's
available to all nodes, but it comes at a price and often must be asynchronously accessed
to achieve good performance.
> * Table/Namespace configuration
> * Users/Authorizations
> * Problem reports (maybe?)
> * System configuration overrides (maybe?)
> Some benefits we'd see from this:
> * Loss of ZooKeeper doesn't lose table configuration and users.
> * Greatly reduce zookeeper watchers (assume watchers=50*num_tables*num_tservers)
> * Consistent updates of table constraints and all other table properties
> The last note is the most important one IMO. The number of test issues alone that we've
had with constraints not being seen on all servers are bound to affect users.

This message was sent by Atlassian JIRA

View raw message