October 2006 - jboss-dev-forums - Jboss List Archives

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Server side HA and failover

by timfox

Off the top of my head, this is more or less what you want to do: Let's say the failed node is node 1, and the fail over node is node 2. The post offices at node 2 will detect that a server has failed (via JGroups). They then look in their node->failover node mapping to determine whether it must take over responsibility for the failed node. node2 is marked as a failover node for node 1 in the mapping so the post offices at node 2 now realise that they must take over responsibility for the failed nodes queues. They then do the following: Take the name map for node 1 and move all the entries to the name map for node 1. Change the names of the queues by suffixing with node id to ensure they are unique. Replace the remote queue stubs with local clustered queues and load them, keep note of the new local clustered queues for the next step. Now iterate through the condition map. For each cluster router on each condition map entry, iterate through the entries in the router. For each one from node 1, replace with the relevant localclusteredQueue from the previous step. Server side fail over should now be complete. The next step will be client reconnection from the client side. When a consumer is reconnected to the queue, just need to look it up with the new name (the old name suffix the node id) and hey presto. There should be no need for any post office signature changes AFAIK. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981039#3981039 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981039

19 years, 5 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Server side HA and failover

by timfox

"clebert.suconic(a)jboss.com" wrote : | Most of the time I'm spending now is looking at the code for how we could couple with this (open the failed ChannelID into a diffferent server), maybe changing Routers and other objects. | If you think you could do this part faster it would be great. Ok, let me have a think about it and I'll try and come up with a more explicit/detailed solution :) View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981028#3981028 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981028

19 years, 5 months

1
0
0 / 0

[Design of JBoss Portal] - Re: LDAP Support expectations

by bvogt

Hi Boleslaw, I'm interested in the "identity responsibilities decisions" you mentioned above from the view of the database usage. Our customer needs role assignments to groups of users in order to easily authorize a new user with a set of portal funtionality. "Provisioning" shall take place as "pull service" at the portal side, which periodically requests the entitlement changes and updates the portals data. Therefore the portals database can/shall be used (in order to have just one server/service). Managing role assignments to groups wthin the portal is a "nice to have" feature, but xml-based mass-changes is required. I would like to provide these features on the basis of your design decisions. Burkhard cf: http://jboss.org/index.html?module=bb&op=viewtopic&t=91902 View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981027#3981027 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981027

19 years, 5 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Server side HA and failover

by clebert.suconic＠jboss.com

Tim wrote : We just need to extend the model slightly to cope with this. Most of the time I'm spending now is looking at the code for how we could couple with this (open the failed ChannelID into a diffferent server), maybe changing Routers and other objects. If you think you could do this part faster it would be great. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981026#3981026 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981026

19 years, 5 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Server side HA and failover

by timfox

"clebert.suconic(a)jboss.com" wrote : | | But anyway, If you have 10s of millions of messages you still need to load these messages anyway. | | Actually, no. If you have 10s of millions of messages then 99.99% of them will probably be paged in storage. You only need to load those few thousand or so that will fit in memory. This is a big difference. "Clebert" wrote : | Tim wrote : Why is this a problem? | | What we already discussed about a PostOffice only accepting one binding per name identified by a String. We could use the nodeID on the ChannelName but on nodeB it will still create a new channel onde nodeB what would require state transfer between one node and another node. | | We just need to extend the model slightly to cope with this. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981023#3981023 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981023

19 years, 5 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Client failover redeliveries discussion

by clebert.suconic＠jboss.com

Ovidiu wrote : To spell this out for you, so you won't go over is again: The current code, that is in the SVN, does not behave this way. This is understandable, it's just a prototype. I was going over the use cases for which we need test cases, so we can make sure the correct behavior is preserved in the future. The correct behavior is "transparently copying the transactional state (the corresponding TxState instance) into the new ResourceManager and send the messages over the new connection when transaction commits." Please note the use of the word "transparent". | The Transaction still has any pending transaction on its buffer when a connection is failed over. The code it's not a prototype IMO, it's one of the layers of the failover and currently I'm not considering the first layer yet. I have started from the second layer as when we started we were not sure what was going to happen with Remoting. I want to finish this layer now before starting on the first layer. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981021#3981021 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981021

19 years, 5 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: Server side HA and failover

by clebert.suconic＠jboss.com

Tim wrote : As already discussed, moving messages is not an option since there may be 10s of millions of messages in each partial queue. I"m just pointing out what we discuseed on the other thread. But anyway, If you have 10s of millions of messages you still need to load these messages anyway. Tim wrote : Why is this a problem? What we already discussed about a PostOffice only accepting one binding per name identified by a String. We could use the nodeID on the ChannelName but on nodeB it will still create a new channel onde nodeB what would require state transfer between one node and another node. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981015#3981015 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981015

19 years, 5 months

1
0
0 / 0

[Design of JBossCache] - Re: Eviction thread behaviour

by manik.surtani＠jboss.com

Setting the eviction thread lock acquisition timeout to 0 will allow the Eviction Thread to process the rest of the queue without blocking on trying to evict stuff that is locked. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981001#3981001 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981001

19 years, 5 months

1
0
0 / 0

[Design of JBossCache] - Re: Eviction thread behaviour

by manik.surtani＠jboss.com

anonymous wrote : | | What I don't understand is still the cause of the problem. I understand you can't reproduce it reading from the Jira, right? | | I couldn't reproduce it because of a timing problem, but I do completely understand the cause of the problem. Consider (LIFO): 1) Eviction queue is close to full, cache region is full. 2) Start a tx 3) Add stuff to the cache 4) Causes older items in the region to be queued for eviction 5) tx reads item in cache, which was queued for eviction 6) Node visited event in 5) not yet received, Eviction Thread attemps to process queue. Waiting on RL in 5) 7) tx attempts to write more stuff, but blocks because this triggers more evictions and the eviction queue is now full. The tx doesn't get a chance to commit and release the RL in 5), because 7) blocks. The eviction thread cannot empty the quete because it is waiting n 5). Deadlock, until lock timeout! View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3981000#3981000 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3981000

19 years, 5 months

1
0
0 / 0

[Design of JBossCache] - Re: Removal of old code in 2.0 release

by manik.surtani＠jboss.com

+1. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=3980998#3980998 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=3980998

19 years, 5 months

1
0
0 / 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

jboss-dev-forums October 2006