Sync from server repo (9e6d6f6f46f)

vertica · May 13, 2024 · 9b9e623 · 9b9e623
1 parent 520dab9
commit 9b9e623
Show file tree

Hide file tree

Showing 34 changed files with 96 additions and 66 deletions.
diff --git a/.vscode/launch.json b/.vscode/launch.json
diff --git a/commands/cmd_add_node.go b/commands/cmd_add_node.go
@@ -71,7 +71,7 @@ Examples:
     --data-path /data --hosts 10.20.30.40 \
     --node-names v_test_db_node0001,v_test_db_node0002
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, dataPathFlag, depotPathFlag,
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, dataPathFlag, depotPathFlag,
 			passwordFlag},
 	)
 

diff --git a/commands/cmd_add_subcluster.go b/commands/cmd_add_subcluster.go
@@ -77,7 +77,7 @@ Examples:
 	--hosts 10.20.30.40,10.20.30.41,10.20.30.42 \
 	--is-primary --control-set-size -1 --new-hosts 10.20.30.43
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, eonModeFlag, passwordFlag,
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, eonModeFlag, passwordFlag,
 			dataPathFlag, depotPathFlag},
 	)
 

diff --git a/commands/cmd_drop_db.go b/commands/cmd_drop_db.go
@@ -63,7 +63,7 @@ Examples:
   vcluster drop_db --db-name test_db \
     --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, catalogPathFlag, dataPathFlag, depotPathFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, catalogPathFlag, dataPathFlag, depotPathFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_install_packages.go b/commands/cmd_install_packages.go
@@ -63,7 +63,7 @@ Examples:
   vcluster install_packages --db-name test_db --force-reinstall \
     --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, passwordFlag, outputFileFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, passwordFlag, outputFileFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_list_all_nodes.go b/commands/cmd_list_all_nodes.go
@@ -62,7 +62,7 @@ Examples:
   vcluster list_allnodes --password testpassword \
     --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, hostsFlag, passwordFlag, catalogPathFlag, configFlag, outputFileFlag},
+		[]string{dbNameFlag, hostsFlag, passwordFlag, ipv6Flag, catalogPathFlag, configFlag, outputFileFlag},
 	)
 
 	return cmd
@@ -77,6 +77,9 @@ func (c *CmdListAllNodes) Parse(inputArgv []string, logger vlog.Printer) error {
 	// reset the value of those options to nil
 	c.ResetUserInputOptions(&c.fetchNodeStateOptions.DatabaseOptions)
 
+	// Set GetVersion to true so that the CLI can retrieve versions for down nodes
+	// by invoking two additional operations: NMAHealth and NMA readCatalogEditor
+	c.fetchNodeStateOptions.GetVersion = true
 	return c.validateParse(logger)
 }
 

diff --git a/commands/cmd_re_ip.go b/commands/cmd_re_ip.go
@@ -67,7 +67,7 @@ Examples:
   vcluster re_ip --db-name test_db --re-ip-file /data/re_ip_map.json \
     --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, hostsFlag, catalogPathFlag, configParamFlag, configFlag},
+		[]string{dbNameFlag, hostsFlag, ipv6Flag, catalogPathFlag, configParamFlag, configFlag},
 	)
 
 	// local flags
@@ -93,7 +93,9 @@ func (c *CmdReIP) setLocalFlags(cmd *cobra.Command) {
 func (c *CmdReIP) Parse(inputArgv []string, logger vlog.Printer) error {
 	c.argv = inputArgv
 	logger.LogArgParse(&c.argv)
-
+	// Set CheckDBRunning to true so that CLI can check running db for Re_IP
+	// Re-IP should only be used for down DB, checking if db is running
+	c.reIPOptions.CheckDBRunning = true
 	return c.validateParse(logger)
 }
 

diff --git a/commands/cmd_remove_node.go b/commands/cmd_remove_node.go
@@ -61,7 +61,7 @@ Examples:
   vcluster db_remove_node --db-name test_db --remove 10.20.30.42 \
     --hosts 10.20.30.40 --data-path /data
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, catalogPathFlag, dataPathFlag, depotPathFlag, passwordFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, catalogPathFlag, dataPathFlag, depotPathFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_remove_subcluster.go b/commands/cmd_remove_subcluster.go
@@ -59,7 +59,7 @@ Examples:
     --hosts 10.20.30.40,10.20.30.41,10.20.30.42 --subcluster sc1 \
     --data-path /data --depot-path /data
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, eonModeFlag, dataPathFlag, depotPathFlag, passwordFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, eonModeFlag, dataPathFlag, depotPathFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_restart_node.go b/commands/cmd_restart_node.go
@@ -75,7 +75,7 @@ Examples:
     --restart v_test_db_node0003=10.20.30.42,v_test_db_node0004=10.20.30.43 \
     --password testpassword --config /opt/vertica/config/vertica_cluster.yaml	
 `,
-		[]string{dbNameFlag, hostsFlag, configFlag, passwordFlag},
+		[]string{dbNameFlag, hostsFlag, ipv6Flag, configFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_revive_db.go b/commands/cmd_revive_db.go
@@ -79,7 +79,7 @@ Examples:
     --ignore-cluster-lease --restore-point-archive db --restore-point-index 1
 
 `,
-		[]string{dbNameFlag, hostsFlag, communalStorageLocationFlag, configFlag, outputFileFlag, configParamFlag},
+		[]string{dbNameFlag, hostsFlag, ipv6Flag, communalStorageLocationFlag, configFlag, outputFileFlag, configParamFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_sandbox.go b/commands/cmd_sandbox.go
@@ -69,7 +69,7 @@ Examples:
   vcluster sandbox_subcluster --subcluster sc1 --sandbox sand \
     --hosts 10.20.30.40,10.20.30.41,10.20.30.42 --db-name test_db
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, passwordFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_scrutinize.go b/commands/cmd_scrutinize.go
@@ -104,7 +104,7 @@ Examples:
   vcluster scrutinize --db-name test_db --db-user dbadmin \
     --password testpassword --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, hostsFlag, configFlag, catalogPathFlag, passwordFlag},
+		[]string{dbNameFlag, hostsFlag, ipv6Flag, configFlag, catalogPathFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_show_restore_points.go b/commands/cmd_show_restore_points.go
@@ -79,7 +79,7 @@ Examples:
     --start-timestamp 2024-03-04 08:32:33.277569 \
     --end-timestamp 2024-03-04 08:32:34.176391
 `,
-		[]string{dbNameFlag, configFlag, passwordFlag, hostsFlag,
+		[]string{dbNameFlag, configFlag, passwordFlag, hostsFlag, ipv6Flag,
 			communalStorageLocationFlag, configParamFlag},
 	)
 

diff --git a/commands/cmd_start_db.go b/commands/cmd_start_db.go
@@ -66,7 +66,7 @@ Examples:
   vcluster start_db --password testpassword \
     --config /opt/vertica/config/vertica_cluster.yaml
 `,
-		[]string{dbNameFlag, hostsFlag, communalStorageLocationFlag,
+		[]string{dbNameFlag, hostsFlag, communalStorageLocationFlag, ipv6Flag,
 			configFlag, catalogPathFlag, passwordFlag, eonModeFlag, configParamFlag},
 	)
 

diff --git a/commands/cmd_start_subcluster.go b/commands/cmd_start_subcluster.go
@@ -56,7 +56,7 @@ Examples:
   vcluster start_subcluster --db-name test_db \
     --hosts 10.20.30.40,10.20.30.41,10.20.30.42 --subcluster sc1
 `,
-		[]string{dbNameFlag, configFlag, hostsFlag, eonModeFlag, passwordFlag},
+		[]string{dbNameFlag, configFlag, hostsFlag, ipv6Flag, eonModeFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_stop_node.go b/commands/cmd_stop_node.go
@@ -53,7 +53,7 @@ Examples:
   vcluster stop_node --db-name test_db --stop-hosts 10.20.30.40,10.20.30.41 \
     --hosts 10.20.30.40,10.20.30.41,10.20.30.42 
 `,
-		[]string{dbNameFlag, hostsFlag, configFlag, passwordFlag},
+		[]string{dbNameFlag, hostsFlag, ipv6Flag, configFlag, passwordFlag},
 	)
 
 	// local flags

diff --git a/commands/cmd_unsandbox.go b/commands/cmd_unsandbox.go
@@ -73,7 +73,7 @@ Examples:
   vcluster unsandbox_subcluster --subcluster sc1 \
     --hosts 10.20.30.40,10.20.30.41,10.20.30.42 --db-name test_db
 `,
-		[]string{dbNameFlag, configFlag, passwordFlag, hostsFlag},
+		[]string{dbNameFlag, configFlag, ipv6Flag, passwordFlag, hostsFlag},
 	)
 
 	// local flags

diff --git a/vclusterops/adapter_pool.go b/vclusterops/adapter_pool.go
@@ -32,20 +32,22 @@ type adapterPool struct {
 }
 
 var (
-	poolInstance *adapterPool
+	poolInstance adapterPool
 	once         sync.Once
 )
 
-// return a singleton instance of the AdapterPool
-func getPoolInstance(logger vlog.Printer) *adapterPool {
+// return a new instance of an adapterPool. The adapterPool cannot be shared
+// between Go routines. Otherwise, they will clobber each other state causing
+// HTTP request errors. It is the callers responsibility to ensure it doesn't
+// get shared.
+func getPoolInstance(logger vlog.Printer) adapterPool {
 	/* if once.Do(f) is called multiple times,
 	 * only the first call will invoke f,
 	 * even if f has a different value in each invocation.
 	 * Reference: https://pkg.go.dev/sync#Once
 	 */
 	once.Do(func() {
-		pool := makeAdapterPool(logger)
-		poolInstance = &pool
+		poolInstance = makeAdapterPool(logger)
 	})
 
 	return poolInstance

diff --git a/vclusterops/cluster_op.go b/vclusterops/cluster_op.go
@@ -365,7 +365,7 @@ func (op *opBase) loadCertsIfNeeded(certs *httpsCerts, findCertsInOptions bool)
 
 	// this step is executed after Prepare() so all http requests should be set up
 	if len(op.clusterHTTPRequest.RequestCollection) == 0 {
-		return fmt.Errorf("[%s] has not set up a http request", op.name)
+		return fmt.Errorf("[%s] clusterHTTPRequest.RequestCollection is empty", op.name)
 	}
 
 	if certs == nil {

diff --git a/vclusterops/coordinator_database.go b/vclusterops/coordinator_database.go
@@ -331,8 +331,9 @@ type VCoordinationNode struct {
 	// empty string if it is not an eon db
 	Subcluster string
 	// empty string if it is not in a sandbox
-	Sandbox string
-	Version string
+	Sandbox       string
+	Version       string
+	IsControlNode bool
 }
 
 func makeVCoordinationNode() VCoordinationNode {

diff --git a/vclusterops/fetch_node_state.go b/vclusterops/fetch_node_state.go
@@ -8,6 +8,10 @@ import (
 
 type VFetchNodeStateOptions struct {
 	DatabaseOptions
+	// retrieve the version for down nodes by invoking two additional
+	// operations: NMAHealth and NMA readCatalogEditor. This is useful
+	// when we cannot get the version for down nodes from a running database
+	GetVersion bool
 }
 
 func VFetchNodeStateOptionsFactory() VFetchNodeStateOptions {
@@ -183,9 +187,13 @@ func (vcc VClusterCommands) produceListAllNodesInstructions(
 		return instructions, err
 	}
 
+	if options.GetVersion {
+		instructions = append(instructions,
+			&nmaHealthOp,
+			&nmaReadVerticaVersionOp)
+	}
+
 	instructions = append(instructions,
-		&nmaHealthOp,
-		&nmaReadVerticaVersionOp,
 		&httpsCheckNodeStateOp,
 	)
 

diff --git a/vclusterops/helpers.go b/vclusterops/helpers.go
@@ -109,6 +109,7 @@ type nodeStateInfo struct {
 	Name             string   `json:"name"`
 	Sandbox          string   `json:"sandbox_name"`
 	Version          string   `json:"build_info"`
+	IsControlNode    bool     `json:"is_control_node"`
 }
 
 func (node *nodeStateInfo) asNodeInfo() (n NodeInfo, err error) {

diff --git a/vclusterops/http_request_dispatcher.go b/vclusterops/http_request_dispatcher.go
@@ -22,7 +22,7 @@ import (
 
 type requestDispatcher struct {
 	opBase
-	pool *adapterPool
+	pool adapterPool
 }
 
 func makeHTTPRequestDispatcher(logger vlog.Printer) requestDispatcher {

diff --git a/vclusterops/https_get_nodes_info_op.go b/vclusterops/https_get_nodes_info_op.go
@@ -146,6 +146,7 @@ func (op *httpsGetNodesInfoOp) processResult(_ *opEngineExecContext) error {
 				vNode.State = node.State
 				vNode.Subcluster = node.Subcluster
 				vNode.Sandbox = node.Sandbox
+				vNode.IsControlNode = node.IsControlNode
 				if node.IsPrimary && node.State == util.NodeUpState {
 					op.vdb.PrimaryUpNodes = append(op.vdb.PrimaryUpNodes, node.Address)
 				}

diff --git a/vclusterops/https_poll_node_state_op.go b/vclusterops/https_poll_node_state_op.go
@@ -26,7 +26,7 @@ import (
 // Timeout set to 30 seconds for each GET /v1/nodes/{node} call.
 // 30 seconds is long enough for normal http request.
 // If this timeout is reached, it might imply that the target IP is unreachable
-const defaultHTTPRequestTimeoutSeconds = 30
+const defaultHTTPSRequestTimeoutSeconds = 30
 const (
 	StartDBCmd CmdType = iota
 	StartNodeCmd
@@ -67,7 +67,7 @@ func makeHTTPSPollNodeStateOpHelper(hosts []string,
 	op.description = fmt.Sprintf("Wait for %d node(s) to come up", len(hosts))
 	op.hosts = hosts
 	op.useHTTPPassword = useHTTPPassword
-	op.httpRequestTimeout = defaultHTTPRequestTimeoutSeconds
+	op.httpRequestTimeout = defaultHTTPSRequestTimeoutSeconds
 	op.checkDown = false // setting default to poll nodes UP
 	err := util.ValidateUsernameAndPassword(op.name, useHTTPPassword, userName)
 	if err != nil {

diff --git a/vclusterops/https_poll_subcluster_node_state_op.go b/vclusterops/https_poll_subcluster_node_state_op.go
@@ -86,7 +86,7 @@ func (op *httpsPollSubclusterNodeStateOp) setupClusterHTTPRequest(hosts []string
 	for _, host := range hosts {
 		httpRequest := hostHTTPRequest{}
 		httpRequest.Method = GetMethod
-		httpRequest.Timeout = defaultHTTPRequestTimeoutSeconds
+		httpRequest.Timeout = defaultHTTPSRequestTimeoutSeconds
 		httpRequest.buildHTTPSEndpoint("nodes/" + host)
 		if op.useHTTPPassword {
 			httpRequest.Password = op.httpsPassword

diff --git a/vclusterops/https_poll_subscription_state_op.go b/vclusterops/https_poll_subscription_state_op.go
@@ -61,7 +61,7 @@ func (op *httpsPollSubscriptionStateOp) setupClusterHTTPRequest(hosts []string)
 	for _, host := range hosts {
 		httpRequest := hostHTTPRequest{}
 		httpRequest.Method = GetMethod
-		httpRequest.Timeout = defaultHTTPRequestTimeoutSeconds
+		httpRequest.Timeout = defaultHTTPSRequestTimeoutSeconds
 		httpRequest.buildHTTPSEndpoint("subscriptions")
 		if op.useHTTPPassword {
 			httpRequest.Password = op.httpsPassword

diff --git a/vclusterops/https_update_node_state_op.go b/vclusterops/https_update_node_state_op.go
@@ -25,6 +25,8 @@ type httpsUpdateNodeStateOp struct {
 	opBase
 	opHTTPSBase
 	vdb *VCoordinationDatabase
+	// The timeout for each http request. Requests will be repeated if timeout hasn't been exceeded.
+	httpRequestTimeout int
 }
 
 func makeHTTPSUpdateNodeStateOp(vdb *VCoordinationDatabase,
@@ -35,9 +37,9 @@ func makeHTTPSUpdateNodeStateOp(vdb *VCoordinationDatabase,
 	op := httpsUpdateNodeStateOp{}
 	op.name = "HTTPSUpdateNodeStateOp"
 	op.description = "Update node state from running database"
-	op.hosts = vdb.HostList
 	op.vdb = vdb
 	op.useHTTPPassword = useHTTPPassword
+	op.httpRequestTimeout = defaultHTTPSRequestTimeoutSeconds
 
 	err := util.ValidateUsernameAndPassword(op.name, useHTTPPassword, userName)
 	if err != nil {
@@ -53,6 +55,7 @@ func (op *httpsUpdateNodeStateOp) setupClusterHTTPRequest(hosts []string) error
 	for _, host := range hosts {
 		httpRequest := hostHTTPRequest{}
 		httpRequest.Method = GetMethod
+		httpRequest.Timeout = op.httpRequestTimeout
 		httpRequest.buildHTTPSEndpoint("nodes/" + host)
 		if op.useHTTPPassword {
 			httpRequest.Password = op.httpsPassword
@@ -65,6 +68,7 @@ func (op *httpsUpdateNodeStateOp) setupClusterHTTPRequest(hosts []string) error
 }
 
 func (op *httpsUpdateNodeStateOp) prepare(execContext *opEngineExecContext) error {
+	op.hosts = op.vdb.HostList
 	execContext.dispatcher.setup(op.hosts)
 
 	return op.setupClusterHTTPRequest(op.hosts)