[Building Sakai] Clustering portal chat
Alex Balleste
alexandre.balleste at udl.cat
Thu Jul 25 12:56:17 PDT 2013
Hi, after some test we enabled neoportal chat in our 3 cluster
production servers. I used the following configuration in
jgroups-config.xml based on JGroups example. Messages flow correctly, no
error messages found in logs about disconnections.
<config>
<UDP
mcast_addr="${jgroups.udp.mcast_addr:224.X.X.X}"
mcast_port="${jgroups.udp.mcast_port:55555}"
discard_incompatible_packets="true"
max_bundle_size="60000"
max_bundle_timeout="30"
ip_ttl="${jgroups.udp.ip_ttl:2}"
enable_bundling="true"
ucast_recv_buf_size="103K"
ucast_send_buf_size="64K"
mcast_recv_buf_size="103K"
mcast_send_buf_size="64K"
thread_pool.enabled="true"
thread_pool.min_threads="1"
thread_pool.max_threads="25"
thread_pool.keep_alive_time="5000"
thread_pool.queue_enabled="false"
thread_pool.queue_max_size="100"
thread_pool.rejection_policy="Run"
oob_thread_pool.enabled="true"
oob_thread_pool.min_threads="1"
oob_thread_pool.max_threads="8"
oob_thread_pool.keep_alive_time="5000"
oob_thread_pool.queue_enabled="false"
oob_thread_pool.queue_max_size="100"
oob_thread_pool.rejection_policy="Run"/>
<PING timeout="2000"
num_initial_members="3"/>
<MERGE2 max_interval="30000"
min_interval="10000"/>
<FD_SOCK/>
<FD timeout="10000" max_tries="5" />
<VERIFY_SUSPECT timeout="1500" />
<pbcast.NAKACK
use_mcast_xmit="false"
retransmit_timeout="300,600,1200,2400,4800"
discard_delivered_msgs="true"/>
<UNICAST timeout="300,600,1200,2400,3600"/>
<pbcast.STABLE stability_delay="1000" desired_avg_gossip="50000"
max_bytes="400000"/>
<pbcast.GMS print_local_addr="true" join_timeout="3000"
view_bundling="true"/>
<FC max_credits="20000000"
min_threshold="0.10"/>
<FRAG2 frag_size="30000" />
<pbcast.STATE_TRANSFER />
</config>
On 07/17/2013 10:44 PM, Omer A Piperdi wrote:
> I think it is JBoss specific.. Here is pbcast.GMS properties
> http://www.jgroups.org/manual/html/protlist.html#d0e6042
>
> Omer
> On 7/17/2013 3:31 PM, Juan José Meroño Sánchez wrote:
>>
>> It seems that one cluster member was shunned. Did you try to
>> configure shun to false? I don't know if this is a valid option for
>> mcast.
>>
>> El 17/07/2013 20:29, "Omer A Piperdi" <omer at rice.edu
>> <mailto:omer at rice.edu>> escribió:
>>
>> Actually, I have to increase the join_timeout in jGroup
>> pbcast.GMS to make it work.. or start two Tomcat at the same time
>> will work also..
>>
>> I would like to know if anyone running portal chat in
>> production?? How are they handling this issue. In case if you
>> need to bring one machine down from cluster and put it back in
>> later, portal chat might not work..
>>
>> Thanks
>> Omer
>> On 7/17/2013 11:50 AM, Omer A Piperdi wrote:
>>> Thanks Adrian.. I am running a program to monitor the jGroup
>>> session multicast traffic and I see both machine talking to ip
>>> address and port that I specify in config. file.
>>>
>>> But I am getting following error in catalina.out and still
>>> cannot chat between the boxes..
>>>
>>> 2013-07-17 10:22:16,464 WARN
>>> OOB-63,owlspacechat,sakaidev-n1-6688 org.jgroups.protocols.UDP -
>>> sakaidev-n1-6688: no physical address for
>>> 7edaa4b2-cf26-cc70-fdd3-e0541bf29883, dropping message
>>> 2013-07-17 10:22:21,907 WARN
>>> OOB-63,owlspacechat,sakaidev-n1-6688
>>> org.jgroups.protocols.pbcast.NAKACK - sakaidev-n1-6688: dropped
>>> message 437 from 7edaa4b2-cf26-cc70-fdd3-e0541bf29883 (sender
>>> not in table [sakaidev-n1-6688]), view=[sakaidev-n1-6688|0]
>>> [sakaidev-n1-6688]
>>>
>>> Thanks,
>>> Omer
>>> On 7/17/2013 4:07 AM, Adrian Fish wrote:
>>>> Alternatively, you can supply params to jgroups via system
>>>> properties.
>>>>
>>>> *-Djgroups.udp.mcast_addr=xxx.xxx.xxx.xxx
>>>> -Djgroups.udp.mcast_port=xx*
>>>> *
>>>> *
>>>> *I got that from here:*
>>>> *
>>>> *
>>>> *https://community.jboss.org/wiki/SystemProps?_sscc=t
>>>> *
>>>> *
>>>> *
>>>> *cheers again,*
>>>> *Adrian.*
>>>>
>>>>
>>>> On 16 July 2013 23:28, Omer A Piperdi <omer at rice.edu
>>>> <mailto:omer at rice.edu>> wrote:
>>>>
>>>> We are trying to set up neo portal chat in a cluster
>>>> environment and I
>>>> am not able to successfully configured. It only works when
>>>> users login
>>>> into same machine, if a user is logged into another
>>>> machine, chat tool
>>>> says "User is offline" even the user is not offline..
>>>>
>>>> Here is what I have done with the settings. Since I want to
>>>> use specific
>>>> address and port for multi-casting, so I created
>>>> jgroups-config.xml file
>>>> under tomcat/sakai folder. But I don't think Sakai is
>>>> loading the
>>>> mcast_addr and mcast_port that I specify in the file.
>>>>
>>>> I get this message during startup..
>>>> -------------------------------------------------------------------
>>>> GMS: address=sakaidev-n1-32110, cluster=owlspacechat, physical
>>>> address=10.64.93.212:33844 <http://10.64.93.212:33844>
>>>> -------------------------------------------------------------------
>>>> 2013-07-16 17:11:22,532 INFO localhost-startStop-1
>>>> org.sakaiproject.portal.chat.entity.PCServiceEntityProvider
>>>> - Portal
>>>> chat is connected on JGroups channel 'owlspacechat'
>>>>
>>>> But the physical address here is the ip address of the
>>>> host, not the
>>>> mcast_addr that I specified in config file..
>>>>
>>>> Where should I put the mcast_address so that both machine
>>>> use the same
>>>> address???
>>>>
>>>> Thanks
>>>> Omer
>>>> _______________________________________________
>>>> sakai-dev mailing list
>>>> sakai-dev at collab.sakaiproject.org
>>>> <mailto:sakai-dev at collab.sakaiproject.org>
>>>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>>>
>>>> TO UNSUBSCRIBE: send email to
>>>> sakai-dev-unsubscribe at collab.sakaiproject.org
>>>> <mailto:sakai-dev-unsubscribe at collab.sakaiproject.org> with
>>>> a subject of "unsubscribe"
>>>>
>>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> sakai-dev mailing list
>>> sakai-dev at collab.sakaiproject.org <mailto:sakai-dev at collab.sakaiproject.org>
>>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>>
>>> TO UNSUBSCRIBE: send email tosakai-dev-unsubscribe at collab.sakaiproject.org <mailto:sakai-dev-unsubscribe at collab.sakaiproject.org> with a subject of "unsubscribe"
>>>
>>> !DSPAM:2294,51e6cb62102191134720216!
>>
>>
>> _______________________________________________
>> sakai-dev mailing list
>> sakai-dev at collab.sakaiproject.org
>> <mailto:sakai-dev at collab.sakaiproject.org>
>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>
>> TO UNSUBSCRIBE: send email to
>> sakai-dev-unsubscribe at collab.sakaiproject.org
>> <mailto:sakai-dev-unsubscribe at collab.sakaiproject.org> with a
>> subject of "unsubscribe"
>>
>> !DSPAM:2294,51e6ff24102195331411829!
>
>
>
> _______________________________________________
> sakai-dev mailing list
> sakai-dev at collab.sakaiproject.org
> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>
> TO UNSUBSCRIBE: send email to sakai-dev-unsubscribe at collab.sakaiproject.org with a subject of "unsubscribe"
--
Alexandre Ballesté Crevillén alexandre.balleste at udl.cat
=====================
University of Lleida
Information and Communication Systems Service
Analyst / Programmer
Twitter: @sinmsinm
Tlf: +34 973 702148
Fax: +34 973 702130
=====================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20130725/2f34602c/attachment.html
More information about the sakai-dev
mailing list