Howdy,
Are the resource pool settings (i.e. maxmemorysize, plannedconcurrency, etc) per node or per cluster?
Resource Pool Settings
Moderator: NorbertKrupa
-
- GURU
- Posts: 527
- Joined: Tue Oct 22, 2013 9:36 pm
- Location: Chicago, IL
- Contact:
Re: Resource Pool Settings
If you take a look at the documentation, you should find your answer in the descriptions of the parameters.
MEMORYSIZE 'sizeUnits'
Amount of memory allocated to the resource pool. Memory allocated to this pool is per node and not across the whole cluster.
PLANNEDCONCURRENCY
This is a cluster-wide maximum and not a per-node limit.
Checkout vertica.tips for more Vertica resources.
Re: Resource Pool Settings
I laugh, because this documentation is actually incorrect:
PLANNEDCONCURRENCY does not set a limit. MAXCONCURRENCY sets a limit. PLANNEDCONCURRENCY provides a guideline for the purpose of budgeting memory for queries but it does not provide any kind of guarantee on concurrency or maximum on concurrency.PLANNEDCONCURRENCY
This is a cluster-wide maximum and not a per-node limit.
Sharon Cutter
Vertica Consultant, Zazz Technologies LLC
Vertica Consultant, Zazz Technologies LLC
Re: Resource Pool Settings
Thanks all! Now I get it
I was a little confused as to why the planned_concurrency column is in the resource_pool_status table which can be broken out by node (node_name column)... I guess the planned_concurrency will be the same for each pool on all nodes.
I was a little confused as to why the planned_concurrency column is in the resource_pool_status table which can be broken out by node (node_name column)... I guess the planned_concurrency will be the same for each pool on all nodes.
-
- GURU
- Posts: 527
- Joined: Tue Oct 22, 2013 9:36 pm
- Location: Chicago, IL
- Contact:
Re: Resource Pool Settings
Let Sarah know!scutter wrote:I laugh, because this documentation is actually incorrect:.
Checkout vertica.tips for more Vertica resources.
Re: Resource Pool Settings
If plannedconcurrency is set to AUTO, then it can result in different values on different nodes in resource_pool_status if the servers aren’t homogeneous - one server having more cores than others. That’s not ideal, but happens sometimes as new hardware is added to an existing cluster without upgrading older nodes.
Norbert, I shall enter a bug report against the documentation
Norbert, I shall enter a bug report against the documentation
Sharon Cutter
Vertica Consultant, Zazz Technologies LLC
Vertica Consultant, Zazz Technologies LLC