[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16814151#comment-16814151 ] Keshav Varma commented on CASSANDRA-7544: - Ah, yes, I was testing with the upstream PyPI releases of the driver. With the custom egg from the repo, things seem to work as expected. Thanks for the pointer! > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Normal > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16812968#comment-16812968 ] Ariel Weisberg commented on CASSANDRA-7544: --- [~keshavdv] there is a checked in driver under lib that should have the argument. Do you have another version of the Cassandra driver in your environment that might be getting in the way? You are right it's not merged into the Datastax python driver. It's blocked on https://datastax-oss.atlassian.net/browse/JAVA-2105 > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Normal > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16812873#comment-16812873 ] Keshav Varma commented on CASSANDRA-7544: - [~aweisberg], I see you added support for this in cqlsh at [https://github.com/apache/cassandra/blame/2d622e05d1576d20c3bf6504cbdaf438a895b4cf/bin/cqlsh.py#L490], but I can't find any corresponding commit in the published python-driver that makes this work. In the meantime, this ends up breaking cqlsh for me on trunk since that keyword argument doesn't exist ([https://github.com/datastax/python-driver/blob/4.x/cassandra/cluster.py#L707]). Is there a different fork or branch that includes the driver side changes that makes this work for you? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Normal > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16440983#comment-16440983 ] Ariel Weisberg commented on CASSANDRA-7544: --- I think it's fine to tweak what that is called. You are right --with-port seems wrong as you would expect it to accept a port number as an argument with the Java builder idiom and maybe also CLI idioms. I'll create a separate ticket for it. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16440600#comment-16440600 ] Stefan Podkowinski commented on CASSANDRA-7544: --- Having both --port and --with-port options for nodetool is slightly confusing. Can we rename --with-port to something like --print-port? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16399020#comment-16399020 ] Jeff Jirsa commented on CASSANDRA-7544: --- No - it's very invasive, and it was written on top of a networking refactor that would be incredibly painful to re-write for 3.11. It would also introduce a lot of risk into 3.11, which is a few releases in and needs to stabilize. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16399015#comment-16399015 ] Chris Remshaw commented on CASSANDRA-7544: -- Is there any possibility of merging this resolution into a 3.x release? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16369714#comment-16369714 ] Ariel Weisberg commented on CASSANDRA-7544: --- Committed ninja fix as [cc436770010034f5e9fbde06b9a2e713ddbffcee|https://github.com/apache/cassandra/commit/cc436770010034f5e9fbde06b9a2e713ddbffcee] > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16369682#comment-16369682 ] Ariel Weisberg commented on CASSANDRA-7544: --- Sure Kurt I will change that. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16369610#comment-16369610 ] Kurt Greaves commented on CASSANDRA-7544: - [~aweisberg] one nit: [this|https://github.com/apache/cassandra/blob/720772033a1f92b422065a307dcacc5f024b9c07/src/java/org/apache/cassandra/service/StorageService.java#L2014] if statement uses a bitwise OR in place of a logical OR. While the behaviour is the same, can you ninja a fix? Just to save us when the condition gets changed in the future and someone doesn't notice. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16357355#comment-16357355 ] Jason Brown commented on CASSANDRA-7544: {{StorageService#getRemovalStatus(boolean)}} is only called from via the MBean interface; thus I'm not sure throwing the {{ConcurrentModificationException}} is really worthwhile. A slightly stale answer (on the order of milliseconds) from this call is good enough for it's intended use. As long as you take the snapshot of {{replicatingNodes }}(as you are doing) that should be reasonable. Otherwise, +1 > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16353329#comment-16353329 ] Dave Brosius commented on CASSANDRA-7544: - also StorageService ~4354 {code:java} for (InetAddressAndPort node : replicatingNodes) { toFormat.add(node.toString(false)); } {code} replicatingNodes is a synchronized set so iterating would seem unsafe. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16349213#comment-16349213 ] Adam Holmberg commented on CASSANDRA-7544: -- {quote}The primary key changed. We don't support schema changes of the primary key so there has to be a new table. {quote} Gotcha. I was focused on the other attributes and looked right past that. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16349197#comment-16349197 ] Ariel Weisberg commented on CASSANDRA-7544: --- bq. I don't think this is accurate. Protocol does not dictate which tables are present, server version does. Driver support isn't merged yet so if we want to use server version to do that we can. bq. On another note, is there a reason this was implemented as a new system table (system.peers_v2), rather than new columns in the existing system.peers? I The primary key changed. We don't support schema changes of the primary key so there has to be a new table. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16349193#comment-16349193 ] Adam Holmberg commented on CASSANDRA-7544: -- Can we also discuss how the node metadata is implemented? {quote}The clients use the protocol version to select the correct system tables when querying metadata. {quote} I don't think this is accurate. Protocol does not dictate which tables are present, server version does. On another note, is there a reason this was implemented as a new system table ({{system.peers_v2}}), rather than new columns in the existing {{system.peers}}? I think it might be cleaner from the client perspective if the existing table is expanded, rather than having to probe or conditionally query different tables. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16349143#comment-16349143 ] Ariel Weisberg commented on CASSANDRA-7544: --- Working on reverting the protocol version changes now. CASSANDRA-14211. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16348921#comment-16348921 ] Ariel Weisberg commented on CASSANDRA-7544: --- bq. ISTM that a discussion on the dev list is warranted. That's a pretty big "side effect" of this patch. It's not clear that it is? It sounds like you know something? I'd be happy to accept a patch > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16348816#comment-16348816 ] Jonathan Ellis commented on CASSANDRA-7544: --- {quote}I'm not sure how we intended that to work. We don't have trunk releases so what is the expectation there from the perspective of clients? {quote} ISTM that a discussion on the dev list is warranted. That's a pretty big "side effect" of this patch. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16345901#comment-16345901 ] Ariel Weisberg commented on CASSANDRA-7544: --- bq. I see that the protocol version is incremented but there are no edits to the native_protocol spec. Oversight? The native protocol didn't change per se. The clients use the protocol version to select the correct system tables when querying metadata. It's definitely something I could include in the spec, although historically the spec is incomplete WRT to the contents of Cassandra system tables. bq. Also, is this the right place to change the default protocol? Shouldn't that be a separate discussion? What is the default protocol? You mean move v5 out of beta? I'm not sure how we intended that to work. We don't have trunk releases so what is the expectation there from the perspective of clients? We could put it in back in beta and update the code elsewhere that relies on the beta functionality to be correct to use the beta version (cqlsh, sstableloader, maybe the hadoop stuff). Then when we move it out of beta we have to remove the use beta flag otherwise we will be releasing utilities that use the next beta version (v6) of the protocol. Seems a bit like the tail wagging the dog, but then end result is similar. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16345873#comment-16345873 ] Jonathan Ellis commented on CASSANDRA-7544: --- I see that the protocol version is incremented but there are no edits to the native_protocol spec. Oversight? Also, is this the right place to change the default protocol? Shouldn't that be a separate discussion? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.0 > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16339614#comment-16339614 ] Jason Brown commented on CASSANDRA-7544: [~aweisberg] and I have been going back and forth about technical details on the PR, and I think we're basically there. I think we can get rid of the {{CompactEndpointSerializationHelper}} specialization for streaming because we don't stream across major versions, and we drop the connection when the streaming version doesn't match (from the internode protocol for streaming), so we won't have a peer who doesn't send the IP addr in the expected format (either ip alone (old style), or ip/port). Once we confirm that, we should be good to commit. I think you'll need to get ccm committed and released, then turn around and commit the dtest and c* patches. I have not reviewed either of the drivers as I think a driver author (one who can commit to the respective repos) should review. However, I am willing to review if a driver author cannot find time in the short term. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16337800#comment-16337800 ] Ariel Weisberg commented on CASSANDRA-7544: --- The latest branches currently under review: dtests: [https://github.com/aweisberg/cassandra-dtest/tree/cassandra-7544-rebased] cassandra: [https://github.com/aweisberg/cassandra/tree/cassandra-7544-rebased3] ccm: [https://github.com/riptano/ccm/pull/654] Java driver: [ JAVA-1388Support clusters with different native and storage ports at each node using CASSANDRA-7544|https://datastax-oss.atlassian.net/browse/JAVA-1388] Python driver: [ PYTHON-700Support clusters with different native and storage ports at each node using CASSANDRA-7544|https://datastax-oss.atlassian.net/browse/PYTHON-700] > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg >Priority: Major > Fix For: 4.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v7.6.3#76005) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16144374#comment-16144374 ] Ariel Weisberg commented on CASSANDRA-7544: --- I think that depends on the 4.0 release date. This was really kind of sort of ready to go in quite a while ago, but we decided to make it depend on Netty and using a single port for both SSL and non-SSL so that we would only use a single port throughout the code base rather than two and transition away from two ports later. Jason says that CASSANDRA-10404 is almost done. I need to do a lot of rebasing. On to of Netty in C*, but also the Java and Python driver, and those haven't been reviewed yet. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 4.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16144187#comment-16144187 ] Jon Haddad commented on CASSANDRA-7544: --- Is this patch destined for 4.0? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 3.11.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15850104#comment-15850104 ] Brandon Williams commented on CASSANDRA-7544: - Cleared the reviewer. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15849445#comment-15849445 ] sankalp kohli commented on CASSANDRA-7544: -- [~brandon.williams] I see that you are assigned the reviewer of this JIRA way back in 2014. Will you be reviewing this or we find someone else? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15849342#comment-15849342 ] Ariel Weisberg commented on CASSANDRA-7544: --- This introduces a client library impacting change for CASSANDRA-11596 that is currently disabled by a flag so that client libraries that claim to support the next version of the protocol continue to work. [~rkuris] how are things going with the Java client? Currently no one has done the Python driver or any of the other drivers. Not sure how we go about integrating changes that require drivers to support a newer version of the protocol. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15849339#comment-15849339 ] Ariel Weisberg commented on CASSANDRA-7544: --- ||Code|utests|dtests|| |[trunk|https://github.com/apache/cassandra/compare/trunk...aweisberg:cassandra-7544-rebased?expand=1]|[utests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-cassandra-7544-rebased-testall/4/]|[dtests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-cassandra-7544-rebased-dtest/6]| [dtest changes|https://github.com/riptano/cassandra-dtest/compare/master...aweisberg:master] > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Ariel Weisberg > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15327996#comment-15327996 ] sankalp kohli commented on CASSANDRA-7544: -- cc [~pmcfadin] > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement >Reporter: Sam Overton >Assignee: Sam Overton > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14640989#comment-14640989 ] Brandon Williams commented on CASSANDRA-7544: - No. > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement > Components: Core >Reporter: Sam Overton >Assignee: Sam Overton > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (CASSANDRA-7544) Allow storage port to be configurable per node
[ https://issues.apache.org/jira/browse/CASSANDRA-7544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14640900#comment-14640900 ] Raju S commented on CASSANDRA-7544: --- so prior versions of 3.X, can we not able to configure different ports for each node in cluster for rpc_port and storage_port? > Allow storage port to be configurable per node > -- > > Key: CASSANDRA-7544 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7544 > Project: Cassandra > Issue Type: Improvement > Components: Core >Reporter: Sam Overton >Assignee: Sam Overton > Fix For: 3.x > > > Currently storage_port must be configured identically on all nodes in a > cluster and it is assumed that this is the case when connecting to a remote > node. > This prevents running in any environment that requires multiple nodes to be > able to bind to the same network interface, such as with many automatic > provisioning/deployment frameworks. > The current solutions seems to be > * use a separate network interface for each node deployed to the same box. > This puts a big requirement on IP allocation at large scale. > * allow multiple clusters to be provisioned from the same resource pool, but > restrict allocation to a maximum of one node per host from each cluster, > assuming each cluster is running on a different storage port. > It would make operations much simpler in these kind of environments if the > environment provisioning the resources could assign the ports to be used when > bringing up a new node on shared hardware. > The changes required would be at least the following: > 1. configure seeds as IP:port instead of just IP > 2. gossip the storage port as part of a node's ApplicationState > 3. refer internally to nodes by hostID instead of IP, since there will be > multiple nodes with the same IP > (1) & (2) are mostly trivial and I already have a patch for these. The bulk > of the work to enable this is (3), and I would structure this as a separate > pre-requisite patch. -- This message was sent by Atlassian JIRA (v6.3.4#6332)