ADDREPLICA and maxShardsPerNode and specifying a node

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

ADDREPLICA and maxShardsPerNode and specifying a node

Erick Erickson
I noticed today that if I create a collection with maxShardsPerNode, I
can freely exceed that number by specifying the "node" parameter on an
ADDREPLICA command.

Is this intended behavior, or should I raise a JIRA? I can argue that
if a user is really going to specify a node, then we should do
whatever they say even if it exceeds maxShardsPerNode. If this is
intended though I'll want to explicitly call that out in ref guide.

Erick

P.S. I'd _really_ like to rename maxShardsPerNode to
maxReplicasPerNode, but it's too late for that.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: ADDREPLICA and maxShardsPerNode and specifying a node

Shai Erera

I think that if addreplica is called, we should assume the user understands what he's doing, and we can safely ignore maxShardsPerNode.

+1 in renaming it. I don't think it's late to do anything. First, it probably is an advanced parameter to be specified, if not expert. Second, we can always deprecate it and add maxReplicasPerNode, support both and remove the former in 2 minor releases  or the next major one.

IMO it's more important to have good API with sensible parameter names, than leave it as that because of back-compat concerns.

Shai

On Jul 9, 2015 8:33 AM, "Erick Erickson" <[hidden email]> wrote:
I noticed today that if I create a collection with maxShardsPerNode, I
can freely exceed that number by specifying the "node" parameter on an
ADDREPLICA command.

Is this intended behavior, or should I raise a JIRA? I can argue that
if a user is really going to specify a node, then we should do
whatever they say even if it exceeds maxShardsPerNode. If this is
intended though I'll want to explicitly call that out in ref guide.

Erick

P.S. I'd _really_ like to rename maxShardsPerNode to
maxReplicasPerNode, but it's too late for that.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: ADDREPLICA and maxShardsPerNode and specifying a node

david.w.smiley@gmail.com
+1 to all Shai said.

On Thu, Jul 9, 2015 at 1:42 AM Shai Erera <[hidden email]> wrote:

I think that if addreplica is called, we should assume the user understands what he's doing, and we can safely ignore maxShardsPerNode.

+1 in renaming it. I don't think it's late to do anything. First, it probably is an advanced parameter to be specified, if not expert. Second, we can always deprecate it and add maxReplicasPerNode, support both and remove the former in 2 minor releases  or the next major one.

IMO it's more important to have good API with sensible parameter names, than leave it as that because of back-compat concerns.

Shai

On Jul 9, 2015 8:33 AM, "Erick Erickson" <[hidden email]> wrote:
I noticed today that if I create a collection with maxShardsPerNode, I
can freely exceed that number by specifying the "node" parameter on an
ADDREPLICA command.

Is this intended behavior, or should I raise a JIRA? I can argue that
if a user is really going to specify a node, then we should do
whatever they say even if it exceeds maxShardsPerNode. If this is
intended though I'll want to explicitly call that out in ref guide.

Erick

P.S. I'd _really_ like to rename maxShardsPerNode to
maxReplicasPerNode, but it's too late for that.

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

--
Lucene/Solr Search Committer, Consultant, Developer, Author, Speaker