Cloud Platform

Shape the future of IBM!

We invite you to shape the future of IBM, including product roadmaps, by submitting ideas that matter to you the most. Here's how it works:

Post your ideas

Start by posting ideas and requests to enhance a product or service. Take a look at ideas others have posted and upvote them if they matter to you,

  1. Post an idea

  2. Upvote ideas that matter most to you

  3. Get feedback from the IBM team to refine your idea

Help IBM prioritize your ideas and requests

The IBM team may need your help to refine the ideas so they may ask for more information or feedback. The offering manager team will then decide if they can begin working on your idea. If they can start during the next development cycle, they will put the idea on the priority list. Each team at IBM works on a different schedule, where some ideas can be implemented right away, others may be placed on a different schedule.

Receive a notification on the decision

Some ideas can be implemented at IBM, while others may not fit within the development plans for the product. In either case, the team will let you know as soon as possible. In some cases, we may be able to find alternatives for ideas which cannot be implemented in a reasonable time.

If you encounter any issues accessing the Ideas portals, please send email describing the issue to ideasibm@us.ibm.com for resolution.

For more information about IBM's Ideas program visit ibm.com/ideas.

Status Under review
Created by Guest
Created on Mar 28, 2017

The Last update time of map entries used by evictors is lost in a shard replication

The following details have been provided in 08953,442,000 and we have been told to raise a RFE to get the issue fixed. This is a HIGH priority issue to ODM, because it implies a bug in the ODM DSI product and a customer has already fall into it few times.

Detail:
For a map with a ttlEvictorType of type LAST_UPDATE_TIME, the last update time of map entries is reset when a map is replicated to another server that has just been started. As a result, if the map in the new server becomes primary, the evictor acts as if the LAST_UPDATE_TIME was the data at which the server started. This is highly visible in ripple start scenarios (when you need to stop and restrart servers one by one for maintainance , for example server update for security fixes and reboot). In that case the LAST_UPDATE_TIME for all the grid entries is lost.

This is seen with Xtreme Scale 8.6.0.8 (+ latest IFIX) - but alo with previous releases.

Idea priority Urgent
Use Case
How to reproduce: A simple scenario to reproduce is: - Install XS - Go to /ObjectGrid/gettingstarted Edit server/config/objectgrid.xml and add ttlEvictorType="LAST_UPDATE_TIME" timeToLive="600" for each backing map - Start a catalog and a server (./startcat then ./startcontainer server1 ) - Start the sample client (in another terminal -- you will need a terminal to start and stop servers) create an entry manually i a 17:00 (for this test it is convenient to put the data as the value) wait for 10 minutes (600 seconds - the TTL of the map) and query the value g a -> The result is 'not found' as expected Then create another entry i a 17:05 wait for 5 minutes start another server ( ./startcontainer server2 ) -> It is important to start it after the entry is created (the defect does not occur for entities that are updated when a replica is alive) wait enough for the replication to take place then stop the server 1 (./stopcontainer server1) 11 minutes after a is created, query it g a -> The value is returned 10 minutes after the server 2 is started, the value is evicted g a -> not found.
RFE ID 102957
RFE URL
RFE Product WebSphere eXtreme Scale