: 4400 | 106120 | 11884
|
: 5 : 10 : 5 |
We have situations where a cluster mate will go down in a cluster with two cluster mates. We usually find out the hard way that a database was not created on the second cluster because the user cannot access their mailbox.
When a user is created, the mailbox is created on their home server.
It would be great if there was a way to ensure that the database was also created on any cluster mates if the user's specific home server is in a cluster. At least a warning or a question dialog box upon the user creation process.
Really what my idea is about though, is creating a method that would allow administrators to initate a synch of all the databases within a cluster. This would be beneficial for application only Domino shops, and mail only domino shops.
Currently, there is the tick box option when creating a new replica that says "show me only cluster members for cluster: <clustername>
This is beneficial, but what I'm talking about is taking it one step further and allowing a process for synching if things weren't done right the first time.
Currently, we can see which databases are not synched by looking at the "Databases by Replica" view in the Cluster directory database (cldbdir.nsf) that is created and replicated between any cluster mates automatically when a server is added to a cluster.
It seems like the framework is there for this process to work, it's deciding where to allow the admin to kick off the process, and generating a button.
Example:
Server A has 30 databases, and is the main mail server in a company.
6 months later, the company sets up Server B and adds it to a cluster with Server A, and creates new replicas for 30 databases from Server A on the new server (let's say the admin overlooks 10 databases).
Over the next 6 months, 30 more mail databases are created, but the admins occassionally forget to create the new replica on Server B.
Additionally, databases become corrupt and are deleted, some users home mail servers move to another cluster, etc. etc. etc. general real world administration stuff.
So someone comes in a does an audit and finds that Server A has 100 databases, 25 that are not on Server B.
Server B has 90 databases, 10 of which are not on Server A.
The process to make sure that all databases on Server A are also on Server B and vice versa can be tedious.
|
1) Charles Robinson5451 (22 Feb 2008)
I do this today with scheduled agents. It would be nice to have some built in process for monitoring and reporting, but it's really not that hard to do yourself.
2) Rob Goudvis6575 (22 Feb 2008)
I agree with Charles, so I demoted this idea.
3) David Killingsworth3970 (22 Feb 2008)
Could you elaborate Chris? What do your scheduled agents do
I realize that it's not that "hard" but it is tedious.
A simple click process would eliminate alot of time spent on this particular housekeeping task. Checking your servers every day to make sure that one database that is one is also on another is not at the top of my priority.
Additionally, I submitted this idea because though, cluster servers don't usually go down, and we have to failover to the other cluster....when it does happen, we want to spend our time working on getting the original cluster member back online, not dealing with angry users who can't access their applications/mail files.
I realize that it's not that "hard" but it is tedious.
A simple click process would eliminate alot of time spent on this particular housekeeping task. Checking your servers every day to make sure that one database that is one is also on another is not at the top of my priority.
Additionally, I submitted this idea because though, cluster servers don't usually go down, and we have to failover to the other cluster....when it does happen, we want to spend our time working on getting the original cluster member back online, not dealing with angry users who can't access their applications/mail files.
4) Charles Robinson5451 (22 Feb 2008)
David, I'll assume you're talking to me (my name is Charles, not Chris). I have an agent that runs at 9 PM that scans the cldbdir.nsf and verifies whether things are in synch. Any database that should be replicated but isn't is created on the other servers at that time.
I was genuinely torn about whether to promote or demote this. I demoted this because it's easy enough to fix that I don't want Lotus spending any development cycles on it. It's a good idea, though, and should eventually be part of the product. It's just not important enough to me right now.
I was genuinely torn about whether to promote or demote this. I demoted this because it's easy enough to fix that I don't want Lotus spending any development cycles on it. It's a good idea, though, and should eventually be part of the product. It's just not important enough to me right now.
5) JP Liggett456 (23 Feb 2008)
Is the agent available open source? Some sort of cluster check facility to notify or create mates is needed to ensure proper fail-over. Without this agent or facility, it's too time consuming to be done on a regular basis. usually the admin finds out after fail-over, and the customers perception of notes suffers. I promote this idea.
6) Randy LeTourneau75 (05 Mar 2008)
Another option from Paul Looney's AdminBlast @ Lotusphere is not automatic, but really easy.
In Files tab of Admin client, right click on folder with mailboxes and select New Replica(s). This will create new replicas for any mailbox that doesn't have one.
In Files tab of Admin client, right click on folder with mailboxes and select New Replica(s). This will create new replicas for any mailbox that doesn't have one.
7) David Killingsworth3970 (18 Mar 2008)
Update: I attended the Lotusphere comes to you in Hong Kong recently and in version 8.01 or 8.5 (I can't remember which) there will be a server cluster feature/option which allows you to do exactly what I described.
You can configure clustering on a server basis instead of a database by database basis. So by putting a server mate into a cluster, it will synch the databases automatically, instead of letting you (or requiring you) to create the replicas manually that will be clustered.
You can configure clustering on a server basis instead of a database by database basis. So by putting a server mate into a cluster, it will synch the databases automatically, instead of letting you (or requiring you) to create the replicas manually that will be clustered.
8) Jean-Yves Riverin371 (28 Mar 2008)
I would be a good idea to be included.
For mr, I do this with my application
{ Link }
If a db is missing, i get an error.
JYR
For mr, I do this with my application
{ Link }
If a db is missing, i get an error.
JYR
Welcome to IdeaJam™
You can run IdeaJam™ in your company. It's easy to install, setup and customize. Your employees, partners and customers will immediately see results.
Use IdeaJam to:
- Collect ideas from employees
- Solicit feedback and suggestions from employees and customers
- Run innovation contests and competitions
- Validate concepts
- Use the power of "crowd-sourcing" to rank ideas and allow the best ideas to rise to the top
IdeaJam™ works with:
- IBM Connections
- IBM Lotus Quickr
- Blogs and Wikis
- Websphere Portal
- Microsoft Sharepoint
- and other applications.
IdeaJam has an extensive set of widgets and API's that allow you to extend and integrate IdeaJam™ with other applications.
Learn more about IdeaJam >>
IdeaJam developed by
