-
- Veeam Vanguard
- Posts: 629
- Liked: 251 times
- Joined: Sep 27, 2011 12:17 pm
- Full Name: Craig Dalrymple
- Location: Scotland
- Contact:
N+1 resource considerations
Can't seem to find the right info for this, so asking here.
Is the management pack smart enough to take N+1 or HA requirements ?
e.g.
I have 3 hosts all 256 , a total of 768GB across the cluser. But realistically I only have 512GB, taking 1 hosts worth of resources for N+1 or HA requirements.
Let's say I look at %memory used metric for a cluster, and see a figure of 60% , is that 60% of 768GB or 60% of 512GB ?
Is the management pack smart enough to take N+1 or HA requirements ?
e.g.
I have 3 hosts all 256 , a total of 768GB across the cluser. But realistically I only have 512GB, taking 1 hosts worth of resources for N+1 or HA requirements.
Let's say I look at %memory used metric for a cluster, and see a figure of 60% , is that 60% of 768GB or 60% of 512GB ?
-
- VP, Product Management
- Posts: 1497
- Liked: 384 times
- Joined: Jan 01, 2006 1:01 am
- Contact:
Re: N+1 resource considerations
Hi Craig,
The % utilization figure is related to the total usage across all hosts in the Cluster. So it would be 60% of 768GB, in your case.
However you can of course set your own threshold for alerting from the monitor for % usage - for example, use an override to create alerting at 60% usage instead of 90%.
Also note that Capacity Planning reports also have a configurable threshold level, so you can forecast resources based on the same 60% threshold.
And you may find the Host Failure Modeling report very useful - that will tell you if you have enough resources to support 1 (or more) host failures in your cluster.
Cheers
Alec
The % utilization figure is related to the total usage across all hosts in the Cluster. So it would be 60% of 768GB, in your case.
However you can of course set your own threshold for alerting from the monitor for % usage - for example, use an override to create alerting at 60% usage instead of 90%.
Also note that Capacity Planning reports also have a configurable threshold level, so you can forecast resources based on the same 60% threshold.
And you may find the Host Failure Modeling report very useful - that will tell you if you have enough resources to support 1 (or more) host failures in your cluster.
Cheers
Alec
-
- Veeam Vanguard
- Posts: 629
- Liked: 251 times
- Joined: Sep 27, 2011 12:17 pm
- Full Name: Craig Dalrymple
- Location: Scotland
- Contact:
Re: N+1 resource considerations
thanks Alec
My next thought would be, is there some way to produce a heat map from the resultant host failure modeling? This more for a quick overview of the state of a cluster.
My next thought would be, is there some way to produce a heat map from the resultant host failure modeling? This more for a quick overview of the state of a cluster.
-
- VP, Product Management
- Posts: 1497
- Liked: 384 times
- Joined: Jan 01, 2006 1:01 am
- Contact:
Re: N+1 resource considerations
Hi Craig,
You can't create heatmaps directly from the modelling report (because heatmaps work on 'live' performance data, they can't access the calculations used in our modeling or trending)
But you could create heatmaps and set your own custom thresholds which factor in the N+1 requirement, as I suggested before for the cpu/memory monitors.
For example - to see a red cluster if memory usage is >N+1 requirement:
Create heatmap -
- Use cluster CPU % to size the tile (there will always be some CPU usage, so the tile will always be visible)
- Then use cluster Memory % to colour the tile.
-Set your colour threshold so that tile is red if memory% is >65%. That would mean your cluster tile would be red at 65 % overall usage which is approx 500GB - in other words, as you approach your N+1 limit of 512GB.
You can't create heatmaps directly from the modelling report (because heatmaps work on 'live' performance data, they can't access the calculations used in our modeling or trending)
But you could create heatmaps and set your own custom thresholds which factor in the N+1 requirement, as I suggested before for the cpu/memory monitors.
For example - to see a red cluster if memory usage is >N+1 requirement:
Create heatmap -
- Use cluster CPU % to size the tile (there will always be some CPU usage, so the tile will always be visible)
- Then use cluster Memory % to colour the tile.
-Set your colour threshold so that tile is red if memory% is >65%. That would mean your cluster tile would be red at 65 % overall usage which is approx 500GB - in other words, as you approach your N+1 limit of 512GB.
-
- Veeam Vanguard
- Posts: 629
- Liked: 251 times
- Joined: Sep 27, 2011 12:17 pm
- Full Name: Craig Dalrymple
- Location: Scotland
- Contact:
Re: N+1 resource considerations
thanks Alec , that's what I thought and how I've set it up, cheers for clarifying and verifying my config
Who is online
Users browsing this forum: No registered users and 3 guests