[JBoss JIRA] (ISPN-8402) Prevent rebalance
by Sebastian Łaskawiec (JIRA)
[ https://issues.jboss.org/browse/ISPN-8402?page=com.atlassian.jira.plugin.... ]
Sebastian Łaskawiec commented on ISPN-8402:
-------------------------------------------
Autoscaling based on custom metrics was implemented in Kubernetes 1.8. I think we should eventually (1.9 or even further down the road) be able to do it, at least to some extend. But frankly, 1.8 is bleeding edge and I haven't had a chance to play with it.
I think it should be possible to instrument OpenShift not to go below certain number of Pods but I really doubt a Pod can veto shutdown process. The only thing we could do is to flush all the data (and transactions) from the node once we receive SIGINT.
> Prevent rebalance
> -----------------
>
> Key: ISPN-8402
> URL: https://issues.jboss.org/browse/ISPN-8402
> Project: Infinispan
> Issue Type: Feature Request
> Components: Cloud Integrations, Core, State Transfer
> Reporter: Sebastian Łaskawiec
> Assignee: Dan Berindei
>
> Both Caching Service and Shared Memory Service require a way to prevent state transfer until the cluster is larger than "target" amount of nodes.
> Note: A thing to consider during the design - we might want to have some timeout here. When we hit it, we might want to do the rebalance regardless to the number of nodes.
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months
[JBoss JIRA] (ISPN-6879) Calculate (and expose) minimum number of nodes for data in Infinispan
by Sebastian Łaskawiec (JIRA)
[ https://issues.jboss.org/browse/ISPN-6879?page=com.atlassian.jira.plugin.... ]
Sebastian Łaskawiec commented on ISPN-6879:
-------------------------------------------
I believe we need to expose it through JMX. This way Hawkular (or Prometheus) will pick it up and we could use it for autoscaling.
Currently users can easily hurt themselves by scaling the cluster down (say from 100 nodes to 1). We need a way to prevent this or send them a warning message.
The idea is to measure how much data we put in the grid (taking number of owners into consideration) and telling - this is the minimum number of nodes you need to operate to store it in memory. If you go below that, you will lose something. Here's an example:
{code}
Number of owners = 1
Each node can store up to 1 GB of data
User dataset size = 10 GB
You can not go below **10** nodes.
{code}
We probably need to keep track of user dataset size as well as how much stuff we can put into the Grid (and this one is very tricky since we probably need to know what is the container capacity but we can check it based on CGroups data (see [here|https://github.com/slaskawi/measure-infinispan/blob/master/docker/me....
> Calculate (and expose) minimum number of nodes for data in Infinispan
> ---------------------------------------------------------------------
>
> Key: ISPN-6879
> URL: https://issues.jboss.org/browse/ISPN-6879
> Project: Infinispan
> Issue Type: Feature Request
> Components: Cloud Integrations, Server
> Reporter: Sebastian Łaskawiec
> Assignee: William Burns
>
> With Kubernetes autoscaling we need to be able to tell what is the minimum amount of nodes necessary for hosting data (probably some sort of size + number of nodes estimation).
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months
[JBoss JIRA] (ISPN-8402) Prevent rebalance
by Dan Berindei (JIRA)
[ https://issues.jboss.org/browse/ISPN-8402?page=com.atlassian.jira.plugin.... ]
Dan Berindei commented on ISPN-8402:
------------------------------------
[~sebastian.laskawiec] this should work for the Caching Service, but it seems dangerous for the Shared Memory Service.
Can we expose an endpoint for OpenShift to change the target number of nodes before shutting down a node and/or query the cluster if it's safe to shut down a node (and we veto the shutdown if the current number of nodes is <= target number of nodes)?
> Prevent rebalance
> -----------------
>
> Key: ISPN-8402
> URL: https://issues.jboss.org/browse/ISPN-8402
> Project: Infinispan
> Issue Type: Feature Request
> Components: Cloud Integrations, Core, State Transfer
> Reporter: Sebastian Łaskawiec
> Assignee: Dan Berindei
>
> Both Caching Service and Shared Memory Service require a way to prevent state transfer until the cluster is larger than "target" amount of nodes.
> Note: A thing to consider during the design - we might want to have some timeout here. When we hit it, we might want to do the rebalance regardless to the number of nodes.
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months
[JBoss JIRA] (ISPN-8402) Prevent rebalance
by Dan Berindei (JIRA)
[ https://issues.jboss.org/browse/ISPN-8402?page=com.atlassian.jira.plugin.... ]
Dan Berindei updated ISPN-8402:
-------------------------------
Status: Open (was: New)
> Prevent rebalance
> -----------------
>
> Key: ISPN-8402
> URL: https://issues.jboss.org/browse/ISPN-8402
> Project: Infinispan
> Issue Type: Feature Request
> Components: Cloud Integrations, Core, State Transfer
> Reporter: Sebastian Łaskawiec
> Assignee: Dan Berindei
>
> Both Caching Service and Shared Memory Service require a way to prevent state transfer until the cluster is larger than "target" amount of nodes.
> Note: A thing to consider during the design - we might want to have some timeout here. When we hit it, we might want to do the rebalance regardless to the number of nodes.
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months
[JBoss JIRA] (ISPN-6879) Calculate (and expose) minimum number of nodes for data in Infinispan
by William Burns (JIRA)
[ https://issues.jboss.org/browse/ISPN-6879?page=com.atlassian.jira.plugin.... ]
William Burns commented on ISPN-6879:
-------------------------------------
It is unclear from the description what we want for this exactly. Do we want this exposed via JMX and then say how many nodes are required to keep up without losing data? Did you want exactly *numOwners* then?
Or is this just a trace message when the cache starts up?
> Calculate (and expose) minimum number of nodes for data in Infinispan
> ---------------------------------------------------------------------
>
> Key: ISPN-6879
> URL: https://issues.jboss.org/browse/ISPN-6879
> Project: Infinispan
> Issue Type: Feature Request
> Components: Cloud Integrations, Server
> Reporter: Sebastian Łaskawiec
> Assignee: William Burns
>
> With Kubernetes autoscaling we need to be able to tell what is the minimum amount of nodes necessary for hosting data (probably some sort of size + number of nodes estimation).
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months
[JBoss JIRA] (ISPN-8376) Add DMR cluster counters management
by Vladimir Blagojevic (JIRA)
[ https://issues.jboss.org/browse/ISPN-8376?page=com.atlassian.jira.plugin.... ]
Vladimir Blagojevic updated ISPN-8376:
--------------------------------------
Description:
We need DMR ops/attributes to manage counters. At the very least:
- list existing counters
- inspect a counter's value
- set/reset a counter value
was:
We need DMR ops to manage counters. At the very least:
- list existing counters
- inspect a counter's value
- set/reset a counter value
> Add DMR cluster counters management
> -----------------------------------
>
> Key: ISPN-8376
> URL: https://issues.jboss.org/browse/ISPN-8376
> Project: Infinispan
> Issue Type: Enhancement
> Components: Clustered Counter, JMX, reporting and management
> Reporter: Tristan Tarrant
> Assignee: Pedro Ruivo
>
> We need DMR ops/attributes to manage counters. At the very least:
> - list existing counters
> - inspect a counter's value
> - set/reset a counter value
--
This message was sent by Atlassian JIRA
(v7.5.0#75005)
6 years, 6 months