4:01 a.m.
David Hotham created JGRP-1496:
----------------------------------
Summary: Views can overtake one another the wrong view is installed
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to GMS.installView,
and then pauses just after the debug trace at line 600 ("installing view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
4:01 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
David Hotham updated JGRP-1496:
-------------------------------
Summary: Views can overtake one another so that the wrong view is installed (was:
Views can overtake one another the wrong view is installed)
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
4:21 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
David Hotham commented on JGRP-1496:
------------------------------------
I've submitted pull request 61.
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
3:52 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
Bela Ban updated JGRP-1496:
---------------------------
Fix Version/s: 3.2
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
Fix For: 3.2
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
3:54 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
Bela Ban commented on JGRP-1496:
--------------------------------
This should not happen if view_ack_collection_timeout is set accordingly
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
Fix For: 3.2
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
1:36 p.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
David Hotham commented on JGRP-1496:
------------------------------------
I agree that it's unlikely (though surely not impossible) for successive views issued
by the same member to arrive in such short time. It's maybe more plausible that
different members might more-or-less simultaneously try to install views that collide in
this way.
I've probably got the focus wrong in the title of this ticket. From my point of view,
much the most important thing about my pull request is that it fixes the race described in
JGRP-1495. Once I thought of the possibility of views overtaking one another: well, that
seemed to generalise the problem and guided the fix. But if you think that we should be
taking some different approach to fixing the JGRP-1495 race, I expect that should be fine
for me.
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
Fix For: 3.2
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.jboss.org/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
5:49 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
Bela Ban commented on JGRP-1496:
--------------------------------
Note to self: view-1 and view-2 are sent by *different* coordinators (merge scenario in
which the coordinator changed, described by JGRP-1495); if they were sent by the *same*
coordinator, everyone would deliver view-1 before view-2.
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
Fix For: 3.2
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
6:24 a.m.
New subject: [JBoss JIRA] (JGRP-1496) Views can overtake one another so that the wrong view is installed
[
https://issues.jboss.org/browse/JGRP-1496?page=com.atlassian.jira.plugin....
]
Bela Ban resolved JGRP-1496.
----------------------------
Resolution: Done
Views can overtake one another so that the wrong view is installed
------------------------------------------------------------------
Key: JGRP-1496
URL: https://issues.jboss.org/browse/JGRP-1496
Project: JGroups
Issue Type: Feature Request
Reporter: David Hotham
Assignee: Bela Ban
Fix For: 3.2
As noted in JGRP-1495; but since I'm now claiming two bugs I should have two
tickets...
- View 1 arrives on Thread Incoming-1. Let's say this thread gets to
GMS.installView, and then pauses just after the debug trace at line 600 ("installing
view")
- View 2 arrives on Thread Incoming-2. It's a newer view, and is fully installed and
notified to everyone.
- Thread Incoming-1 picks up, and completes the installation of View 1
Now everyone has View 1, but the correct view should be View 2.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
4452
days inactive
4490
days old
7 comments
2 participants
participants (2)
-
Bela Ban (JIRA) -
David Hotham (JIRA)