hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrey Stepachev (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-13647) Default value for hbase.client.operation.timeout is too high
Date Fri, 29 May 2015 20:03:17 GMT

    [ https://issues.apache.org/jira/browse/HBASE-13647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14565312#comment-14565312
] 

Andrey Stepachev commented on HBASE-13647:
------------------------------------------

[~enis]
{quote} bq. But we have the same situation with rpc timeout. Should we make an issue to do
something with other timeouts and review them.
RPC timeout is for a single RPC. operation timeout is for the operation with retrying RPCs.
We should not default op timeout to be the operation timeout since it might mean only a single
try, and no retries.
{quote}
Sure. It is overused constant. It seems that it is better to introduce independent value for
region rpc timeout and client operation.
{quote} bq. Say, we can add warnings, or fix them (with some common method like pullUpToZkTimeout)?
Warnings won't cut it. It should be ok to auto set the op timeout by looking at the zk timeout.
If we want (by default) a table operation to ride over RS crashes is taking into account MTTR:
detection + recovery + assignment. Worst case detection is zk session timeout. For recovery
and assignment there is no easy way to infer that. So the default op timeout should be >
zk session timeout IMO.
{quote}
Currently this constant is used for admin operations and endpoint operations.
Would it be better to have separate timeouts? Am I wrong to think that endpoint invocation
should be treated as simple rpc (hense having small rpc comparable timeouts) and operations
like HBaseAdmin operations, which sure can be infinite by default.

> Default value for hbase.client.operation.timeout is too high
> ------------------------------------------------------------
>
>                 Key: HBASE-13647
>                 URL: https://issues.apache.org/jira/browse/HBASE-13647
>             Project: HBase
>          Issue Type: Bug
>    Affects Versions: 2.0.0, 1.0.1, 0.98.13, 1.2.0, 1.1.1
>            Reporter: Andrey Stepachev
>            Assignee: Andrey Stepachev
>            Priority: Critical
>             Fix For: 2.0.0, 0.98.13, 1.0.2, 1.2.0, 1.1.1
>
>         Attachments: HBASE-13647.patch
>
>
> Default value for hbase.client.operation.timeout is too high, it is LONG.Max.
> That value will block any service calls to coprocessor endpoints indefinitely.
> Should we introduce better default value for that?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message