You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@qpid.apache.org by "Robbie Gemmell (JIRA)" <ji...@apache.org> on 2012/05/31 13:22:23 UTC
[jira] [Comment Edited] (QPID-4033) add a connection id to the
session dispatcher thread name to avoid name overloading
[ https://issues.apache.org/jira/browse/QPID-4033?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13286483#comment-13286483 ]
Robbie Gemmell edited comment on QPID-4033 at 5/31/12 11:22 AM:
----------------------------------------------------------------
No, I just mean something that allows differentiating between the same channel numbers in use on different connections, as opposed to something that really identifies the TCP connection.
Host:port has the potential to be very long, and has additional issues such as changing due to failover occurring etc etc, which would require updating the thread names after the fact. I'm not really looking to get into that, so much as provide a means to separate one connection from another, wherever it may be connected. In this case I just added (missed the commit?) a long value that is incremented with each new connection made, and is logged (at debug, could make it info?) along with the connection URL provided (which includes user, and any host:port combinations provided for failover) to allow at least some path for further identification.
The new names are currently e.g Dispatcher-Channel-1-Conn-1. I guess Channel (and now Conn) is a little redundant I guess, I just left the qualifiers because it makes it easier to distinguish what each number is what without looking up the code. "Dispatcher-1-Conn-1" instead perhaps?
was (Author: gemmellr):
No, I just mean something that allows differentiating between the same channel numbers in use on different connections, as opposed to something that really identifies the TCP connection.
Host:port has the potential to be very long, and has additional issues such as changing due to failover occurring etc etc, which would require updating the thread names after the fact. I'm not really looking to get into that, so much as provide a means to separate one connection from another, wherever it may be connected. In this case I just added (missed the commit?) a long value that is incremented with each new connection made, and is logged (at debug) along with the connection URL provided (which includes user, and any host:port combinations provided for failover) to allow at least some path for further identification.
The new names are currently e.g Dispatcher-Channel-1-Conn-1. I guess Channel (and now Conn) is a little redundant I guess, I just left the qualifiers because it makes it easier to distinguish what each number is what without looking up the code. "Dispatcher-1-Conn-1" instead perhaps?
> add a connection id to the session dispatcher thread name to avoid name overloading
> -----------------------------------------------------------------------------------
>
> Key: QPID-4033
> URL: https://issues.apache.org/jira/browse/QPID-4033
> Project: Qpid
> Issue Type: Improvement
> Components: Java Client
> Reporter: Robbie Gemmell
> Assignee: Robbie Gemmell
> Priority: Minor
> Fix For: 0.17
>
>
> The session dispatcher thread is currently called "Dispatcher-Channel-<channel number>", this leads to name overloading if there is mroe than one connection present as the channel numbers are not unique on each connection. We should add a connection id number to allow differentiating the names of dispatcher threads on different connections.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@qpid.apache.org
For additional commands, e-mail: dev-help@qpid.apache.org