[DOCU-8] Add clustering documentation Created: 23/Jan/05  Updated: 21/Jan/08  Resolved: 10/May/05

Status: Closed
Project: Documentation
Component/s: None
Affects Version/s: None
Fix Version/s: None

Type: Bug Priority: Major
Reporter: Angus Mezick Assignee: Magnolia International
Resolution: Fixed Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: XML File server.xml    
Template:
Acceptance criteria:
Empty
Bug DoR:
[ ]* Steps to reproduce, expected, and actual results filled
[ ]* Affected version filled
Date of First Response:

 Description   

I installed author on 8081
I installed two public instances into 8082 and 8083. All instances server pages.
I got changes on 8081 to propagate to 8082. Then I added 8083 into the server.xml and create a 0002 node under subscribers/SubscriberCofing/ in the AdminCentral config panel. The only difference between 0001 and 0002 is that the address node is localhost:8083 instead of localhost. Please improve the documentation to describe how I should get this setup working. I will attach my server.xml.



 Comments   
Comment by Angus Mezick [ 23/Jan/05 ]

Here is how I have configured the server.xml for my 3 instances.

Comment by Angus Mezick [ 23/Jan/05 ]

Ok. I have this running now. I did three things between the last test and now.

1) changed the cache port.
2) Built all the jars from the source and installed them. (maybe this was a bug fix) That build_jars.xml is REALLY nice!!
3) restarted the server.

I still have some confusing with the documentation of the activation.
1) do I have to name them 0001, 0002, 0003, ....?
2) How does this 0001 relate to all the other 0001's?
3) Can I activate only one server in a cluster (Testing vs production)
4) Maybe this is the answer to #3: How would I set up a 3 stage deploy? Maybe I am just waiting for workflows.

I think all this stems from the fact that I am still trying to figure out how to get this system integrated with our dev/qa/production environment. Nothing gets on the dev server (that can have a bug) that doesn't first get cleanly built from the source to the qa server for testing. When we do the QA build, we also create production jars from that code so that we are sure what those jars contain and no label sillyness has taken place.

Comment by Boris Kraft [ 24/Jan/05 ]

1) do I have to name them 0001, 0002, 0003, ....?

No, name them anything you like.

2) How does this 0001 relate to all the other 0001's?

doesn't

3) Can I activate only one server in a cluster (Testing vs production)

Activation goes to all subscribers at once

4) Maybe this is the answer to #3: How would I set up a 3 stage deploy? Maybe I am just waiting for workflows.

author -> test -> live

First you activate from author to test, then from test to live.

Comment by Michael Aemisegger [ 24/Jan/05 ]

A boolean property isActive might simplify administration. Instead of adding and deleting subscribers nodes you simply could change the isActive property.

Comment by Sameer Charles [ 24/Jan/05 ]

I agree, will implement this soon

Comment by Sameer Charles [ 10/May/05 ]

this is already implemented by fabrizio

Generated at Mon Feb 12 01:05:01 CET 2024 using Jira 9.4.2#940002-sha1:46d1a51de284217efdcb32434eab47a99af2938b.