... - Cluster Setup
... - Shard state, leader and followers
... - Entity Owner, candidates and successors
-... - Kill and Start Member
+... - Kill, Stop and Start Member
... - Isolate and Rejoin Member
... - Run Commands On Member
... - REST requests and checks on Members
Resource ${CURDIR}/SSHKeywords.robot
Resource ${CURDIR}/TemplatedRequests.robot # for Get_As_Json_From_Uri
Resource ${CURDIR}/Utils.robot # for Run_Command_On_Controller
+Resource ../variables/Variables.robot
*** Variables ***
${ENTITY_OWNER_URI} restconf/operational/entity-owners:entity-owners
+${GC_LOG_PATH} ${KARAF_HOME}/data/log
${JAVA_HOME} ${EMPTY} # releng/builder scripts should provide correct value
${JOLOKIA_CONF_SHARD_MANAGER_URI} jolokia/read/org.opendaylight.controller:Category=ShardManager,name=shard-manager-config,type=DistributedConfigDatastore
${JOLOKIA_OPER_SHARD_MANAGER_URI} jolokia/read/org.opendaylight.controller:Category=ShardManager,name=shard-manager-operational,type=DistributedOperationalDatastore
+${JOLOKIA_CONFIG_LOCAL_SHARDS_URI} jolokia/read/org.opendaylight.controller:type=DistributedConfigDatastore,Category=ShardManager,name=shard-manager-config/LocalShards
+${JOLOKIA_OPER_LOCAL_SHARDS_URI} jolokia/read/org.opendaylight.controller:type=DistributedOperationalDatastore,Category=ShardManager,name=shard-manager-operational/LocalShards
${JOLOKIA_READ_URI} jolokia/read/org.opendaylight.controller
-${KARAF_HOME} ${WORKSPACE}${/}${BUNDLEFOLDER} # TODO: Migrate to Variables.robot
-@{ODL_DEFAULT_DATA_PATHS} tmp/ data/ cache/ snapshots/ journal/ etc/opendaylight/current/
+# Bug 9044 workaround: delete etc/host.key before restart.
+@{ODL_DEFAULT_DATA_PATHS} tmp/ data/ cache/ snapshots/ journal/ etc/opendaylight/current/ etc/host.key
${RESTCONF_MODULES_DIR} ${CURDIR}/../variables/restconf/modules
+${SINGLETON_NETCONF_DEVICE_ID_PREFIX_OLD} /odl-general-entity:entity[odl-general-entity:name='KeyedInstanceIdentifier{targetType=interface org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.topology.Node, path=[org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.NetworkTopology, org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.Topology[key=TopologyKey [_topologyId=Uri [_value=topology-netconf]]], org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.topology.Node[key=NodeKey [_nodeId=Uri [_value=
+${SINGLETON_NETCONF_DEVICE_ID_SUFFIX_OLD} ]]]]}']
+${SINGLETON_NETCONF_DEVICE_ID_PREFIX} /odl-general-entity:entity[odl-general-entity:name='KeyedInstanceIdentifier{targetType=interface org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.topology.Node, path=[org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.NetworkTopology, org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.Topology[key=TopologyKey{_topologyId=Uri{_value=topology-netconf}}], org.opendaylight.yang.gen.v1.urn.tbd.params.xml.ns.yang.network.topology.rev131021.network.topology.topology.Node[key=NodeKey{_nodeId=Uri{_value=
+${SINGLETON_NETCONF_DEVICE_ID_SUFFIX} }}]]}']
+${SINGLETON_BGPCEP_DEVICE_ID_PREFIX} /odl-general-entity:entity[odl-general-entity:name='
+${SINGLETON_BGPCEP_DEVICE_ID_SUFFIX} -service-group']
+${SINGLETON_SXP_DEVICE_ID_PREFIX} /odl-general-entity:entity[odl-general-entity:name='
+${SINGLETON_SXP_DEVICE_ID_SUFFIX} ']
+${SINGLETON_ELECTION_ENTITY_TYPE} org.opendaylight.mdsal.ServiceEntityType
+${SINGLETON_CHANGE_OWNERSHIP_ENTITY_TYPE} org.opendaylight.mdsal.AsyncServiceCloseEntityType
+${NODE_ROLE_INDEX_START} 1
+${NODE_START_COMMAND} ${KARAF_HOME}/bin/start
+${NODE_STOP_COMMAND} ${KARAF_HOME}/bin/stop
+${NODE_KARAF_COUNT_COMMAND} ps axf | grep org.apache.karaf | grep -v grep | wc -l
+${NODE_KILL_COMMAND} ps axf | grep org.apache.karaf | grep -v grep | awk '{print \"kill -9 \" $1}' | sh
+${NODE_FREEZE_COMMAND} ps axf | grep org.apache.karaf | grep -v grep | awk '{print \"kill -STOP \" $1}' | sh
+${NODE_UNFREEZE_COMMAND} ps axf | grep org.apache.karaf | grep -v grep | awk '{print \"kill -CONT \" $1}' | sh
*** Keywords ***
ClusterManagement_Setup
+ [Arguments] ${http_timeout}=${DEFAULT_TIMEOUT_HTTP} ${http_retries}=0
[Documentation] Detect repeated call, or detect number of members and initialize derived suite variables.
+ ... Http sessions are created with parameters to not waste time when ODL is no accepting connections properly.
# Avoid multiple initialization by several downstream libraries.
${already_done} = BuiltIn.Get_Variable_Value \${ClusterManagement__has_setup_run} False
BuiltIn.Return_From_Keyword_If ${already_done}
${cluster_size} = BuiltIn.Get_Variable_Value \${NUM_ODL_SYSTEM} 1
${status} ${possibly_int_of_members} = BuiltIn.Run_Keyword_And_Ignore_Error BuiltIn.Convert_To_Integer ${cluster_size}
${int_of_members} = BuiltIn.Set_Variable_If '${status}' != 'PASS' ${1} ${possibly_int_of_members}
- ClusterManagement__Compute_Derived_Variables int_of_members=${int_of_members}
+ ClusterManagement__Compute_Derived_Variables int_of_members=${int_of_members} http_timeout=${http_timeout} http_retries=${http_retries}
Check_Cluster_Is_In_Sync
[Arguments] ${member_index_list}=${EMPTY}
[Documentation] Fail if no-sync is detected on a member from list (or any).
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list} # usually: 1, 2, 3.
\ ${status} = Get_Sync_Status_Of_Member member_index=${index}
\ BuiltIn.Continue_For_Loop_If 'True' == '${status}'
\ Get_Leader_And_Followers_For_Shard shard_name=${shard_name} shard_type=${shard_type} validate=True member_index_list=${member_index_list} verify_restconf=${verify_restconf}
Get_Leader_And_Followers_For_Shard
- [Arguments] ${shard_name}=default ${shard_type}=operational ${validate}=True ${member_index_list}=${EMPTY} ${verify_restconf}=True
+ [Arguments] ${shard_name}=default ${shard_type}=operational ${validate}=True ${member_index_list}=${EMPTY} ${verify_restconf}=True ${http_timeout}=${EMPTY}
[Documentation] Get role lists, validate there is one leader, return the leader and list of followers.
... Optionally, issue GET to a simple restconf URL to make sure subsequent operations will not encounter 503.
${leader_list} ${follower_list} = Get_State_Info_For_Shard shard_name=${shard_name} shard_type=${shard_type} validate=True member_index_list=${member_index_list}
- ... verify_restconf=${verify_restconf}
+ ... verify_restconf=${verify_restconf} http_timeout=${http_timeout}
${leader_count} = BuiltIn.Get_Length ${leader_list}
BuiltIn.Run_Keyword_If ${leader_count} < 1 BuiltIn.Fail No leader found.
BuiltIn.Length_Should_Be ${leader_list} ${1} Too many Leaders.
[Return] ${leader} ${follower_list}
Get_State_Info_For_Shard
- [Arguments] ${shard_name}=default ${shard_type}=operational ${validate}=False ${member_index_list}=${EMPTY} ${verify_restconf}=False
+ [Arguments] ${shard_name}=default ${shard_type}=operational ${validate}=False ${member_index_list}=${EMPTY} ${verify_restconf}=False ${http_timeout}=${EMPTY}
[Documentation] Return lists of Leader and Follower member indices from a given member index list
... (or from the full list if empty). If \${shard_type} is not 'config', 'operational' is assumed.
... If \${validate}, Fail if raft state is not Leader or Follower (for example on Candidate).
... The biggest difference from Get_Leader_And_Followers_For_Shard
... is that no check on number of Leaders is performed.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
Collections.Sort_List ${index_list} # to guarantee return values are also sorted lists
# TODO: Support alternative capitalization of 'config'?
${ds_type} = BuiltIn.Set_Variable_If '${shard_type}' != 'config' operational config
${follower_list} = BuiltIn.Create_List
: FOR ${index} IN @{index_list} # usually: 1, 2, 3.
\ ${raft_state} = Get_Raft_State_Of_Shard_At_Member shard_name=${shard_name} shard_type=${ds_type} member_index=${index} verify_restconf=${verify_restconf}
+ \ ... http_timeout=${http_timeout}
\ BuiltIn.Run_Keyword_If 'Follower' == '${raft_state}' Collections.Append_To_List ${follower_list} ${index}
\ ... ELSE IF 'Leader' == '${raft_state}' Collections.Append_To_List ${leader_list} ${index}
\ ... ELSE IF ${validate} BuiltIn.Fail Unrecognized Raft state: ${raft_state}
[Return] ${leader_list} ${follower_list}
Get_Raft_State_Of_Shard_At_Member
- [Arguments] ${shard_name} ${shard_type} ${member_index} ${verify_restconf}=False
+ [Arguments] ${shard_name} ${shard_type} ${member_index} ${verify_restconf}=False ${http_timeout}=${EMPTY}
[Documentation] Send request to Jolokia on indexed member, return extracted Raft status.
... Optionally, check restconf works.
+ ${raft_state} = Get_Raft_Property_From_Shard_Member RaftState ${shard_name} ${shard_type} ${member_index} verify_restconf=${verify_restconf}
+ ... http_timeout=${http_timeout}
+ [Return] ${raft_state}
+
+Get_Raft_Property_From_Shard_Member
+ [Arguments] ${property} ${shard_name} ${shard_type} ${member_index} ${verify_restconf}=False ${http_timeout}=${EMPTY}
+ [Documentation] Send request to Jolokia on indexed member, return extracted Raft property.
+ ... Optionally, check restconf works.
${session} = Resolve_Http_Session_For_Member member_index=${member_index}
# TODO: Does the used URI tend to generate large data which floods log.html?
- BuiltIn.Run_Keyword_If ${verify_restconf} TemplatedRequests.Get_As_Json_Templated session=${session} folder=${RESTCONF_MODULES_DIR} verify=False
+ BuiltIn.Run_Keyword_If ${verify_restconf} TemplatedRequests.Get_As_Json_Templated session=${session} folder=${RESTCONF_MODULES_DIR} verify=False http_timeout=${http_timeout}
${type_class} = Resolve_Shard_Type_Class shard_type=${shard_type}
- ${uri} = BuiltIn.Set_Variable ${JOLOKIA_READ_URI}:Category=Shards,name=member-${member_index}-shard-${shard_name}-${shard_type},type=${type_class}
- ${data_text} = TemplatedRequests.Get_As_Json_From_Uri uri=${uri} session=${session}
+ ${cluster_index} = Evaluate ${member_index}+${NODE_ROLE_INDEX_START}-1
+ ${uri} = BuiltIn.Set_Variable ${JOLOKIA_READ_URI}:Category=Shards,name=member-${cluster_index}-shard-${shard_name}-${shard_type},type=${type_class}
+ ${data_text} = TemplatedRequests.Get_As_Json_From_Uri uri=${uri} session=${session} http_timeout=${http_timeout}
${data_object} = RequestsLibrary.To_Json ${data_text}
${value} = Collections.Get_From_Dictionary ${data_object} value
- ${raft_state} = Collections.Get_From_Dictionary ${value} RaftState
- [Return] ${raft_state}
+ ${raft_property} = Collections.Get_From_Dictionary ${value} ${property}
+ [Return] ${raft_property}
+
+Verify_Shard_Leader_Elected
+ [Arguments] ${shard_name} ${shard_type} ${new_elected} ${old_leader} ${member_index_list}=${EMPTY} ${verify_restconf}=True
+ [Documentation] Verify new leader was elected or remained the same. Bool paramter ${new_elected} indicates if
+ ... new leader is elected or should remained the same as ${old_leader}
+ ${leader} ${followers}= Get_Leader_And_Followers_For_Shard shard_name=${shard_name} shard_type=${shard_type} member_index_list=${member_index_list} verify_restconf=${verify_restconf}
+ BuiltIn.Run_Keyword_If ${new_elected} BuiltIn.Should_Not_Be_Equal_As_Numbers ${old_leader} ${leader}
+ BuiltIn.Run_Keyword_Unless ${new_elected} BuiltIn.Should_Be_Equal_As_numbers ${old_leader} ${leader}
+ BuiltIn.Return_From_Keyword ${leader} ${followers}
Verify_Owner_And_Successors_For_Device
- [Arguments] ${device_name} ${device_type} ${member_index} ${candidate_list}=${EMPTY}
+ [Arguments] ${device_name} ${device_type} ${member_index} ${candidate_list}=${EMPTY} ${after_stop}=False
[Documentation] Returns the owner and successors for the SB device ${device_name} of type ${device_type}. Request is sent to member ${member_index}.
... For Boron and beyond, candidates are not removed on node down or isolation,
... so this keyword expects candidates to be all members from Boron on.
... ${candidate_list} minus owner is returned as ${successor list}.
... Users can still use Get_Owner_And_Successors_For_Device if they are interested in downed candidates,
... or for testing heterogeneous clusters.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${candidate_list}
+ ${index_list} = List_Indices_Or_All given_list=${candidate_list}
${owner} ${successor_list} = Get_Owner_And_Successors_For_Device device_name=${device_name} device_type=${device_type} member_index=${member_index}
Collections.List_Should_Contain_Value ${index_list} ${owner} Owner ${owner} is not in candidate list ${index_list}
- ${expected_candidate_list_origin} = CompareStream.Set_Variable_If_At_Least_Boron ${ClusterManagement__member_index_list} ${index_list}
+ # In Beryllium or after stopping an instance, the removed instance does not show in the candidate list.
+ ${expected_candidate_list_origin} = BuiltIn.Set_Variable_If ${after_stop} ${index_list} ${ClusterManagement__member_index_list}
# We do not want to manipulate either origin list.
${expected_successor_list} = BuiltIn.Create_List @{expected_candidate_list_origin}
Collections.Remove_Values_From_List ${expected_successor_list} ${owner}
[Return] ${owner} ${index_list}
Get_Owner_And_Successors_For_Device
- [Arguments] ${device_name} ${device_type} ${member_index}
+ [Arguments] ${device_name} ${device_type} ${member_index} ${http_timeout}=${EMPTY}
[Documentation] Returns the owner and a list of successors for the SB device ${device_name} of type ${device_type}. Request is sent to member ${member_index}.
... Successors are those device candidates not elected as owner. The list of successors = (list of candidates) - (owner).
... The returned successor list is sorted numerically.
... Note that "candidate list" definition currently differs between Beryllium and Boron.
... Use Verify_Owner_And_Successors_For_Device if you want the older semantics (inaccessible nodes not present in the list).
- ${owner} ${candidate_list} = Get_Owner_And_Candidates_For_Device device_name=${device_name} device_type=${device_type} member_index=${member_index}
+ # TODO: somewhere to introduce ${DEFAULT_RESTCONF_DATASTORE_TIMEOUT}. Value may depend on protocol (ask vs tell) and perhaps stream.
+ ${owner} ${candidate_list} = Get_Owner_And_Candidates_For_Device device_name=${device_name} device_type=${device_type} member_index=${member_index} http_timeout=${http_timeout}
${successor_list} = BuiltIn.Create_List @{candidate_list} # Copy operation is not required, but new variable name requires a line anyway.
Collections.Remove_Values_From_List ${successor_list} ${owner}
[Return] ${owner} ${successor_list}
-Get_Owner_And_Candidates_For_Device
- [Arguments] ${device_name} ${device_type} ${member_index}
+Get_Owner_And_Candidates_For_Device_Old
+ [Arguments] ${device_name} ${device_type} ${member_index} ${http_timeout}=${EMPTY}
[Documentation] Returns the owner and a list of candidates for the SB device ${device_name} of type ${device_type}. Request is sent to member ${member_index}.
... Candidates are all members that register to own a device, so the list of candiates includes the owner.
... The returned candidate list is sorted numerically.
... It is recommended to use Get_Owner_And_Successors_For_Device instead of this keyword, see documentation there.
BuiltIn.Comment TODO: Can this implementation be changed to call Get_Owner_And_Candidates_For_Type_And_Id?
${session} = Resolve_Http_Session_For_Member member_index=${member_index}
- ${data} = TemplatedRequests.Get_As_Json_From_Uri uri=${ENTITY_OWNER_URI} session=${session}
+ ${data} = TemplatedRequests.Get_As_Json_From_Uri uri=${ENTITY_OWNER_URI} session=${session} http_timeout=${http_timeout}
${candidate_list} = BuiltIn.Create_List
${entity_type} = BuiltIn.Set_Variable_If '${device_type}' == 'netconf' netconf-node/${device_name} ${device_type}
${clear_data} = BuiltIn.Run_Keyword_If '${device_type}' == 'openflow' or '${device_type}' == 'netconf' Extract_OpenFlow_Device_Data ${data}
${json} = RequestsLibrary.To_Json ${clear_data}
${entity_type_list} = Collections.Get_From_Dictionary &{json}[entity-owners] entity-type
${entity_type_index} = Utils.Get_Index_From_List_Of_Dictionaries ${entity_type_list} type ${entity_type}
- BuiltIn.Should_Not_Be_Equal ${entity_type_index} -1 No Entity Owner found for ${device_type}
+ BuiltIn.Should_Not_Be_Equal_As_Integers ${entity_type_index} -1 No Entity Owner found for ${device_type}
${entity_list} = Collections.Get_From_Dictionary @{entity_type_list}[${entity_type_index}] entity
${entity_index} = Utils.Get_Index_From_List_Of_Dictionaries ${entity_list} id ${device_name}
- BuiltIn.Should Not Be Equal ${entity_index} -1 Device ${device_name} not found in Entity Owner ${device_type}
+ BuiltIn.Should_Not_Be_Equal_As_Integers ${entity_index} -1 Device ${device_name} not found in Entity Owner ${device_type}
${entity_owner} = Collections.Get_From_Dictionary @{entity_list}[${entity_index}] owner
BuiltIn.Should_Not_Be_Empty ${entity_owner} No owner found for ${device_name}
${owner} = String.Replace_String ${entity_owner} member- ${EMPTY}
Collections.Sort_List ${candidate_list}
[Return] ${owner} ${candidate_list}
+Get_Owner_And_Candidates_For_Device_Singleton
+ [Arguments] ${device_name} ${device_type} ${member_index} ${http_timeout}=${EMPTY}
+ [Documentation] Returns the owner and a list of candidates for the SB device ${device_name} of type ${device_type}. Request is sent to member ${member_index}.
+ ... Parsing method is selected by device type
+ ... Separate kw for every supported device type must be defined
+ BuiltIn.Keyword_Should_Exist Get_Owner_And_Candidates_For_Device_Singleton_${device_type}
+ BuiltIn.Run_Keyword_And_Return Get_Owner_And_Candidates_For_Device_Singleton_${device_type} ${device_name} ${member_index} http_timeout=${http_timeout}
+
+Get_Owner_And_Candidates_For_Device_Singleton_Netconf
+ [Arguments] ${device_name} ${member_index} ${http_timeout}=${EMPTY}
+ [Documentation] Returns the owner and a list of candidates for the SB device ${device_name} of type netconf. Request is sent to member ${member_index}.
+ ... Parsing method is set as netconf (using netconf device id prefix and suffix)
+ # Get election entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_ELECTION_ENTITY_TYPE}
+ ${id} = CompareStream.Set_Variable_If_At_Least_Fluorine ${SINGLETON_NETCONF_DEVICE_ID_PREFIX}${device_name}${SINGLETON_NETCONF_DEVICE_ID_SUFFIX} ${SINGLETON_NETCONF_DEVICE_ID_PREFIX_OLD}${device_name}${SINGLETON_NETCONF_DEVICE_ID_SUFFIX_OLD}
+ ${owner_1} ${candidate_list_1} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Get change ownership entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_CHANGE_OWNERSHIP_ENTITY_TYPE}
+ ${id} = CompareStream.Set_Variable_If_At_Least_Fluorine ${SINGLETON_NETCONF_DEVICE_ID_PREFIX}${device_name}${SINGLETON_NETCONF_DEVICE_ID_SUFFIX} ${SINGLETON_NETCONF_DEVICE_ID_PREFIX_OLD}${device_name}${SINGLETON_NETCONF_DEVICE_ID_SUFFIX_OLD}
+ ${owner_2} ${candidate_list_2} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Owners must be same, if not, there is still some election or change ownership in progress
+ BuiltIn.Should_Be_Equal_As_Integers ${owner_1} ${owner_2} Owners for device ${device_name} are not same
+ [Return] ${owner_1} ${candidate_list_1}
+
+Get_Owner_And_Candidates_For_Device_Singleton_Bgpcep
+ [Arguments] ${device_name} ${member_index} ${http_timeout}=${EMPTY}
+ [Documentation] Returns the owner and a list of candidates for the SB device ${device_name}. Request is sent to member ${member_index}.
+ # Get election entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_ELECTION_ENTITY_TYPE}
+ ${id} = BuiltIn.Set_Variable ${SINGLETON_BGPCEP_DEVICE_ID_PREFIX}${device_name}${SINGLETON_BGPCEP_DEVICE_ID_SUFFIX}
+ ${owner_1} ${candidate_list_1} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Get change ownership entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_CHANGE_OWNERSHIP_ENTITY_TYPE}
+ ${id} = BuiltIn.Set_Variable ${SINGLETON_BGPCEP_DEVICE_ID_PREFIX}${device_name}${SINGLETON_BGPCEP_DEVICE_ID_SUFFIX}
+ ${owner_2} ${candidate_list_2} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Owners must be same, if not, there is still some election or change ownership in progress
+ BuiltIn.Should_Be_Equal_As_Integers ${owner_1} ${owner_2} Owners for device ${device_name} are not same
+ [Return] ${owner_1} ${candidate_list_1}
+
+Get_Owner_And_Candidates_For_Device_Singleton_Sxp
+ [Arguments] ${device_name} ${member_index} ${http_timeout}=${EMPTY}
+ [Documentation] Returns the owner and a list of candidates for the SB device ${device_name}. Request is sent to member ${member_index}.
+ # Get election entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_ELECTION_ENTITY_TYPE}
+ ${id} = BuiltIn.Set_Variable ${SINGLETON_SXP_DEVICE_ID_PREFIX}${device_name}${SINGLETON_SXP_DEVICE_ID_SUFFIX}
+ ${owner_1} ${candidate_list_1} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Get change ownership entity type results
+ ${type} = BuiltIn.Set_Variable ${SINGLETON_CHANGE_OWNERSHIP_ENTITY_TYPE}
+ ${id} = BuiltIn.Set_Variable ${SINGLETON_SXP_DEVICE_ID_PREFIX}${device_name}${SINGLETON_SXP_DEVICE_ID_SUFFIX}
+ ${owner_2} ${candidate_list_2} = Get_Owner_And_Candidates_For_Type_And_Id ${type} ${id} ${member_index} http_timeout=${http_timeout}
+ # Owners must be same, if not, there is still some election or change ownership in progress
+ BuiltIn.Should_Be_Equal_As_Integers ${owner_1} ${owner_2} Owners for device ${device_name} are not same
+ [Return] ${owner_1} ${candidate_list_1}
+
+Get_Owner_And_Candidates_For_Device
+ [Arguments] ${device_name} ${device_type} ${member_index} ${http_timeout}=${EMPTY}
+ [Documentation] Returns the owner and a list of candidates for the SB device ${device_name} of type ${device_type}. Request is sent to member ${member_index}.
+ ... If parsing as singleton failed, kw try to parse data in old way (without singleton).
+ ... Candidates are all members that register to own a device, so the list of candiates includes the owner.
+ ... The returned candidate list is sorted numerically.
+ ... Note that "candidate list" definition currently differs between Beryllium and Boron.
+ ... It is recommended to use Get_Owner_And_Successors_For_Device instead of this keyword, see documentation there.
+ # Try singleton
+ ${status} ${results} = BuiltIn.Run_Keyword_And_Ignore_Error Get_Owner_And_Candidates_For_Device_Singleton device_name=${device_name} device_type=${device_type} member_index=${member_index}
+ ... http_timeout=${http_timeout}
+ BuiltIn.Return_From_Keyword_If "${status}"=="PASS" ${results}
+ # If singleton failed, try parsing in old way
+ ${status} ${results} = BuiltIn.Run_Keyword_And_Ignore_Error Get_Owner_And_Candidates_For_Device_Old device_name=${device_name} device_type=${device_type} member_index=${member_index}
+ ... http_timeout=${http_timeout}
+ # previous 3 lines (BuilIn.Return.., # If singleton..., ${status}....) could be deleted when old way will not be supported anymore
+ BuiltIn.Run_Keyword_If '${status}'=='FAIL' BuiltIn.Fail Could not parse owner and candidates for device ${device_name}
+ [Return] @{results}
+
+Check_Old_Owner_Stays_Elected_For_Device
+ [Arguments] ${device_name} ${device_type} ${old_owner} ${node_to_ask} ${http_timeout}=${EMPTY}
+ [Documentation] Verify the owner remain the same as ${old_owner}
+ ${owner} ${candidates} = Get_Owner_And_Candidates_For_Device ${device_name} ${device_type} ${node_to_ask} http_timeout=${http_timeout}
+ BuiltIn.Should_Be_Equal_As_numbers ${old_owner} ${owner}
+ BuiltIn.Return_From_Keyword ${owner} ${candidates}
+
+Check_New_Owner_Got_Elected_For_Device
+ [Arguments] ${device_name} ${device_type} ${old_owner} ${node_to_ask} ${http_timeout}=${EMPTY}
+ [Documentation] Verify new owner was elected comparing to ${old_owner}
+ ${owner} ${candidates} = Get_Owner_And_Candidates_For_Device ${device_name} ${device_type} ${node_to_ask} http_timeout=${http_timeout}
+ BuiltIn.Should_Not_Be_Equal_As_Numbers ${old_owner} ${owner}
+ BuiltIn.Return_From_Keyword ${owner} ${candidates}
+
Get_Owner_And_Candidates_For_Type_And_Id
- [Arguments] ${type} ${id} ${member_index} ${require_candidate_list}=${EMPTY}
+ [Arguments] ${type} ${id} ${member_index} ${require_candidate_list}=${EMPTY} ${http_timeout}=${EMPTY}
[Documentation] Returns the owner and a list of candidates for entity specified by ${type} and ${id}
... Request is sent to member ${member_index}.
... Candidates are all members that register to own a device, so the list of candiates includes the owner.
... namely you cannot use \${EMPTY} to stand for "full list" in this keyword.
BuiltIn.Comment TODO: Find a way to unify and deduplicate code blocks in Get_Owner_And_Candidates_* keywords.
${session} = Resolve_Http_Session_For_Member member_index=${member_index}
- ${data} = TemplatedRequests.Get_As_Json_From_Uri uri=${ENTITY_OWNER_URI} session=${session}
+ ${data} = TemplatedRequests.Get_As_Json_From_Uri uri=${ENTITY_OWNER_URI} session=${session} http_timeout=${http_timeout}
${candidate_list} = BuiltIn.Create_List
${json} = RequestsLibrary.To_Json ${data}
${entity_type_list} = Collections.Get_From_Dictionary &{json}[entity-owners] entity-type
${entity_type_index} = Utils.Get_Index_From_List_Of_Dictionaries ${entity_type_list} type ${type}
- BuiltIn.Should_Not_Be_Equal ${entity_type_index} -1 No Entity Owner found for ${type}
+ BuiltIn.Should_Not_Be_Equal_As_Integers ${entity_type_index} -1 No Entity Owner found for ${type}
${entity_list} = Collections.Get_From_Dictionary @{entity_type_list}[${entity_type_index}] entity
${entity_index} = Utils.Get_Index_From_List_Of_Dictionaries ${entity_list} id ${id}
- BuiltIn.Should Not Be Equal ${entity_index} -1 Id ${id} not found in Entity Owner ${type}
+ BuiltIn.Should Not_Be_Equal_As_Integers ${entity_index} -1 Id ${id} not found in Entity Owner ${type}
${entity_owner} = Collections.Get_From_Dictionary @{entity_list}[${entity_index}] owner
BuiltIn.Should_Not_Be_Empty ${entity_owner} No owner found for type=${type} id=${id}
${owner} = String.Replace_String ${entity_owner} member- ${EMPTY}
Extract_OpenFlow_Device_Data
[Arguments] ${data}
[Documentation] Remove superfluous OpenFlow device data from Entity Owner printout.
- ${clear_data} = BuiltIn.Run Keyword If '${ODL_STREAM}' != 'beryllium' and '${ODL_OF_PLUGIN}' == 'lithium' String.Replace_String ${data} org.opendaylight.mdsal.ServiceEntityType openflow
- ... ELSE BuiltIn.Set_Variable ${data}
+ ${clear_data} = String.Replace_String ${data} org.opendaylight.mdsal.ServiceEntityType openflow
${clear_data} = String.Replace_String ${clear_data} /odl-general-entity:entity[odl-general-entity:name=' ${EMPTY}
${clear_data} = String.Replace_String ${clear_data} /general-entity:entity[general-entity:name=' ${EMPTY}
${clear_data} = String.Replace_String ${clear_data} '] ${EMPTY}
[Documentation] Convenience keyword that kills the specified member of the cluster.
... The KW will return a list of available members: \${updated index_list}=\${original_index_list}-\${member}
${index_list} = ClusterManagement__Build_List ${member}
+ ${member_ip} = Return_Member_IP ${member}
+ KarafKeywords.Log_Message_To_Controller_Karaf Killing ODL${member} ${member_ip}
${updated_index_list} = Kill_Members_From_List_Or_All ${index_list} ${original_index_list} ${confirm}
[Return] ${updated_index_list}
[Documentation] If the list is empty, kill all ODL instances. Otherwise, kill members based on \${kill_index_list}
... If \${confirm} is True, sleep 1 second and verify killed instances are not there anymore.
... The KW will return a list of available members: \${updated index_list}=\${original_index_list}-\${member_index_list}
- ${kill_index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${original_index_list}
- ${command} = BuiltIn.Set_Variable ps axf | grep karaf | grep -v grep | awk '{print \"kill -9 \" $1}' | sh
- Run_Bash_Command_On_List_Or_All command=${command} member_index_list=${member_index_list}
+ ${kill_index_list} = List_Indices_Or_All given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${original_index_list}
+ Run_Bash_Command_On_List_Or_All command=${NODE_KILL_COMMAND} member_index_list=${member_index_list}
${updated_index_list} = BuiltIn.Create_List @{index_list}
Collections.Remove_Values_From_List ${updated_index_list} @{kill_index_list}
BuiltIn.Return_From_Keyword_If not ${confirm} ${updated_index_list}
BuiltIn.Sleep 1s Kill -9 closes open files, which may take longer than ssh overhead, but not long enough to warrant WUKS.
: FOR ${index} IN @{kill_index_list}
\ Verify_Karaf_Is_Not_Running_On_Member member_index=${index}
+ Run_Bash_Command_On_List_Or_All command=netstat -pnatu | grep 2550
+ [Return] ${updated_index_list}
+
+Stop_Single_Member
+ [Arguments] ${member} ${original_index_list}=${EMPTY} ${confirm}=True ${msg}=${EMPTY}
+ [Documentation] Convenience keyword that stops the specified member of the cluster.
+ ... The KW will return a list of available members: \${updated index_list}=\${original_index_list}-\${member}
+ ${index_list} = ClusterManagement__Build_List ${member}
+ ${member_ip} = Return_Member_IP ${member}
+ ${msg} = Builtin.Set Variable If "${msg}" == "${EMPTY}" Stopping ODL${member} ${member_ip} Stopping ODL${member} ${member_ip}, ${msg}
+ KarafKeywords.Log_Message_To_Controller_Karaf ${msg}
+ ${updated_index_list} = Stop_Members_From_List_Or_All ${index_list} ${original_index_list} ${confirm}
+ [Return] ${updated_index_list}
+
+Stop_Members_From_List_Or_All
+ [Arguments] ${member_index_list}=${EMPTY} ${original_index_list}=${EMPTY} ${confirm}=True ${timeout}=240s
+ [Documentation] If the list is empty, stops all ODL instances. Otherwise stop members based on \${stop_index_list}
+ ... If \${confirm} is True, verify stopped instances are not there anymore.
+ ... The KW will return a list of available members: \${updated index_list}=\${original_index_list}-\${member_index_list}
+ ${stop_index_list} = List_Indices_Or_All given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${original_index_list}
+ Run_Bash_Command_On_List_Or_All command=${NODE_STOP_COMMAND} member_index_list=${member_index_list}
+ ${updated_index_list} = BuiltIn.Create_List @{index_list}
+ Collections.Remove_Values_From_List ${updated_index_list} @{stop_index_list}
+ BuiltIn.Return_From_Keyword_If not ${confirm} ${updated_index_list}
+ : FOR ${index} IN @{stop_index_list}
+ \ BuiltIn.Wait Until Keyword Succeeds ${timeout} 2s Verify_Karaf_Is_Not_Running_On_Member member_index=${index}
+ Run_Bash_Command_On_List_Or_All command=netstat -pnatu | grep 2550
[Return] ${updated_index_list}
Start_Single_Member
- [Arguments] ${member} ${wait_for_sync}=True ${timeout}=300s
+ [Arguments] ${member} ${wait_for_sync}=True ${timeout}=300s ${msg}=${EMPTY} ${check_system_status}=True ${verify_restconf}=True
+ ... ${service_list}=${EMPTY_LIST}
[Documentation] Convenience keyword that starts the specified member of the cluster.
${index_list} = ClusterManagement__Build_List ${member}
- Start_Members_From_List_Or_All ${index_list} ${wait_for_sync} ${timeout}
+ ${member_ip} = Return_Member_IP ${member}
+ ${msg} = Builtin.Set Variable If "${msg}" == "${EMPTY}" Starting ODL${member} ${member_ip} Starting ODL${member} ${member_ip}, ${msg}
+ KarafKeywords.Log_Message_To_Controller_Karaf ${msg}
+ Start_Members_From_List_Or_All ${index_list} ${wait_for_sync} ${timeout} check_system_status=${check_system_status} verify_restconf=${verify_restconf} service_list=${service_list}
Start_Members_From_List_Or_All
- [Arguments] ${member_index_list}=${EMPTY} ${wait_for_sync}=True ${timeout}=300s ${karaf_home}=${KARAF_HOME} ${export_java_home}=${JAVA_HOME}
+ [Arguments] ${member_index_list}=${EMPTY} ${wait_for_sync}=True ${timeout}=300s ${karaf_home}=${EMPTY} ${export_java_home}=${EMPTY} ${gc_log_dir}=${EMPTY}
+ ... ${check_system_status}=True ${verify_restconf}=True ${service_list}=${EMPTY_LIST}
[Documentation] If the list is empty, start all cluster members. Otherwise, start members based on present indices.
... If ${wait_for_sync}, wait for cluster sync on listed members.
... Optionally karaf_home can be overriden. Optionally specific JAVA_HOME is used for starting.
- ${base_command} = BuiltIn.Set_Variable ${karaf_home}/bin/start
- ${command} = BuiltIn.Set_Variable_If "${export_java_home}" export JAVA_HOME="${export_java_home}"; ${base_command} ${base_command}
+ ... Garbage collection is unconditionally logged to files. TODO: Make that reasonable conditional?
+ ${base_command} = BuiltIn.Set_Variable_If """${karaf_home}""" != "" ${karaf_home}/bin/start ${NODE_START_COMMAND}
+ ${command} = BuiltIn.Set_Variable_If """${export_java_home}""" != "" export JAVA_HOME="${export_java_home}"; ${base_command} ${base_command}
+ ${epoch} = DateTime.Get_Current_Date time_zone=UTC result_format=epoch exclude_millis=False
+ ${gc_filepath} = BuiltIn.Set_Variable_If """${karaf_home}""" != "" ${karaf_home}/data/log/gc_${epoch}.log ${GC_LOG_PATH}/gc_${epoch}.log
+ ${gc_options} = BuiltIn.Set_Variable_If "docker" not in """${node_start_command}""" -XX:+PrintGCDetails -XX:+PrintGCDateStamps -Xloggc:${gc_filepath} ${EMPTY}
+ Run_Bash_Command_On_List_Or_All command=${command} ${gc_options} member_index_list=${member_index_list}
+ BuiltIn.Wait_Until_Keyword_Succeeds ${timeout} 10s Verify_Members_Are_Ready ${member_index_list} ${wait_for_sync} ${verify_restconf}
+ ... ${check_system_status} ${service_list}
+ [Teardown] Run_Bash_Command_On_List_Or_All command=netstat -pnatu | grep 2550
+
+Verify_Members_Are_Ready
+ [Arguments] ${member_index_list} ${verify_cluster_sync} ${verify_restconf} ${verify_system_status} ${service_list}
+ [Documentation] Verifies the specified readiness conditions for the given listed members after startup.
+ ... If ${verify_cluster_sync}, verifies the datastores have synced with the rest of the cluster.
+ ... If ${verify_restconf}, verifies RESTCONF is available.
+ ... If ${verify_system_status}, verifies the system services are OPERATIONAL.
+ BuiltIn.Run_Keyword_If ${verify_cluster_sync} Check_Cluster_Is_In_Sync ${member_index_list}
+ BuiltIn.Run_Keyword_If ${verify_restconf} Verify_Restconf_Is_Available ${member_index_list}
+ # for backward compatibility, some consumers might not be passing @{service_list}, but since we can't set a list to a default
+ # value, we need to check here if it's empty in order to skip the check which would throw an error
+ BuiltIn.Run_Keyword_If ${verify_system_status} and ("${service_list}" != "[[]]") ClusterManagement.Check Status Of Services Is OPERATIONAL @{service_list}
+
+Verify_Restconf_Is_Available
+ [Arguments] ${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
+ : FOR ${index} IN @{index_list}
+ \ ${session} = Resolve_Http_Session_For_Member member_index=${index}
+ \ TemplatedRequests.Get_As_Json_Templated session=${session} folder=${RESTCONF_MODULES_DIR} verify=False
+
+Freeze_Single_Member
+ [Arguments] ${member}
+ [Documentation] Convenience keyword that stops the specified member of the cluster by freezing the jvm.
+ ${index_list} = ClusterManagement__Build_List ${member}
+ Freeze_Or_Unfreeze_Members_From_List_Or_All ${NODE_FREEZE_COMMAND} ${index_list}
+
+Unfreeze_Single_Member
+ [Arguments] ${member} ${wait_for_sync}=True ${timeout}=60s
+ [Documentation] Convenience keyword that "continues" the specified member of the cluster by unfreezing the jvm.
+ ${index_list} = ClusterManagement__Build_List ${member}
+ Freeze_Or_Unfreeze_Members_From_List_Or_All ${NODE_UNFREEZE_COMMAND} ${index_list}
+ BuiltIn.Wait_Until_Keyword_Succeeds ${timeout} 10s Check_Cluster_Is_In_Sync
+
+Freeze_Or_Unfreeze_Members_From_List_Or_All
+ [Arguments] ${command} ${member_index_list}=${EMPTY}
+ [Documentation] If the list is empty, stops/runs all ODL instances. Otherwise stop/run members based on \${stop_index_list}
+ ... For command parameter only ${NODE_FREEZE_COMMAND} and ${NODE_UNFREEZE_COMMAND} should be used
+ ${freeze_index_list} = List_Indices_Or_All given_list=${member_index_list}
Run_Bash_Command_On_List_Or_All command=${command} member_index_list=${member_index_list}
- BuiltIn.Return_From_Keyword_If not ${wait_for_sync}
- BuiltIn.Wait_Until_Keyword_Succeeds ${timeout} 10s Check_Cluster_Is_In_Sync member_index_list=${member_index_list}
- # TODO: Do we also want to check Shard Leaders here?
-Clean_Journals_And_Snapshots_On_List_Or_All
+Clean_Journals_Data_And_Snapshots_On_List_Or_All
[Arguments] ${member_index_list}=${EMPTY} ${karaf_home}=${KARAF_HOME}
[Documentation] Delete journal and snapshots directories on every node listed (or all).
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
- ${command} = Set Variable rm -rf "${karaf_home}/journal" "${karaf_home}/snapshots"
+ ... BEWARE: If only a subset of members is cleaned, this causes RetiredGenerationException in Carbon after the affected node re-start.
+ ... See https://bugs.opendaylight.org/show_bug.cgi?id=8138
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
+ ${command} = Set Variable rm -rf "${karaf_home}/journal" "${karaf_home}/snapshots" "${karaf_home}/data"
: FOR ${index} IN @{index_list} # usually: 1, 2, 3.
\ Run_Bash_Command_On_Member command=${command} member_index=${index}
Count_Running_Karafs_On_Member
[Arguments] ${member_index}
[Documentation] Remotely execute grep for karaf process, return count as string.
- ${command} = BuiltIn.Set_Variable ps axf | grep karaf | grep -v grep | wc -l
+ ${command} = BuiltIn.Set_Variable ${NODE_KARAF_COUNT_COMMAND}
${count} = Run_Bash_Command_On_Member command=${command} member_index=${member_index}
[Return] ${count}
Isolate_Member_From_List_Or_All
- [Arguments] ${isolate_member_index} ${member_index_list}=${EMPTY}
+ [Arguments] ${isolate_member_index} ${member_index_list}=${EMPTY} ${protocol}=all ${port}=${EMPTY}
[Documentation] If the list is empty, isolate member from all ODL instances. Otherwise, isolate member based on present indices.
... The KW will return a list of available members: \${updated index_list}=\${member_index_list}-\${isolate_member_index}
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
${source} = Collections.Get_From_Dictionary ${ClusterManagement__index_to_ip_mapping} ${isolate_member_index}
+ ${dport} = BuiltIn.Set_Variable_If '${port}' != '${EMPTY}' --dport ${port} ${EMPTY}
: FOR ${index} IN @{index_list}
\ ${destination} = Collections.Get_From_Dictionary ${ClusterManagement__index_to_ip_mapping} ${index}
- \ ${command} = BuiltIn.Set_Variable sudo /sbin/iptables -I OUTPUT -p all --source ${source} --destination ${destination} -j DROP
+ \ ${command} = BuiltIn.Set_Variable sudo /sbin/iptables -I OUTPUT -p ${protocol} ${dport} --source ${source} --destination ${destination} -j DROP
\ BuiltIn.Run_Keyword_If "${index}" != "${isolate_member_index}" Run_Bash_Command_On_Member command=${command} member_index=${isolate_member_index}
${command} = BuiltIn.Set_Variable sudo /sbin/iptables -L -n
${output} = Run_Bash_Command_On_Member command=${command} member_index=${isolate_member_index}
[Return] ${updated_index_list}
Rejoin_Member_From_List_Or_All
- [Arguments] ${rejoin_member_index} ${member_index_list}=${EMPTY}
+ [Arguments] ${rejoin_member_index} ${member_index_list}=${EMPTY} ${protocol}=all ${port}=${EMPTY} ${timeout}=60s
[Documentation] If the list is empty, rejoin member from all ODL instances. Otherwise, rejoin member based on present indices.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
${source} = Collections.Get_From_Dictionary ${ClusterManagement__index_to_ip_mapping} ${rejoin_member_index}
+ ${dport} = BuiltIn.Set_Variable_If '${port}' != '${EMPTY}' --dport ${port} ${EMPTY}
: FOR ${index} IN @{index_list}
\ ${destination} = Collections.Get_From_Dictionary ${ClusterManagement__index_to_ip_mapping} ${index}
- \ ${command} = BuiltIn.Set_Variable sudo /sbin/iptables -D OUTPUT -p all --source ${source} --destination ${destination} -j DROP
+ \ ${command} = BuiltIn.Set_Variable sudo /sbin/iptables -D OUTPUT -p ${protocol} ${dport} --source ${source} --destination ${destination} -j DROP
\ BuiltIn.Run_Keyword_If "${index}" != "${rejoin_member_index}" Run_Bash_Command_On_Member command=${command} member_index=${rejoin_member_index}
${command} = BuiltIn.Set_Variable sudo /sbin/iptables -L -n
${output} = Run_Bash_Command_On_Member command=${command} member_index=${rejoin_member_index}
BuiltIn.Log ${output}
+ BuiltIn.Wait_Until_Keyword_Succeeds ${timeout} 10s Check_Cluster_Is_In_Sync
Flush_Iptables_From_List_Or_All
[Arguments] ${member_index_list}=${EMPTY}
${command} = BuiltIn.Set_Variable sudo iptables -v -F
${output} = Run_Bash_Command_On_List_Or_All command=${command} member_index_list=${member_index_list}
+Check_Bash_Command_On_List_Or_All
+ [Arguments] ${command} ${member_index_list}=${EMPTY} ${return_success_only}=False ${log_on_success}=True ${log_on_failure}=True ${stderr_must_be_empty}=True
+ [Documentation] Cycle through indices (or all), run bash command on each, using temporary SSH session and restoring the previously active one.
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
+ : FOR ${index} IN @{index_list}
+ \ Check_Bash_Command_On_Member command=${command} member_index=${index} return_success_only=${return_success_only} log_on_success=${log_on_success} log_on_failure=${log_on_failure}
+ \ ... stderr_must_be_empty=${stderr_must_be_empty}
+
+Check_Bash_Command_On_Member
+ [Arguments] ${command} ${member_index} ${return_success_only}=False ${log_on_success}=True ${log_on_failure}=True ${stderr_must_be_empty}=True
+ [Documentation] Open SSH session, call SSHKeywords.Execute_Command_Passes, close session, restore previously active session and return output.
+ BuiltIn.Run_Keyword_And_Return SSHKeywords.Run_Keyword_Preserve_Connection Check_Unsafely_Bash_Command_On_Member ${command} ${member_index} return_success_only=${return_success_only} log_on_success=${log_on_success}
+ ... log_on_failure=${log_on_failure} stderr_must_be_empty=${stderr_must_be_empty}
+
+Check_Unsafely_Bash_Command_On_Member
+ [Arguments] ${command} ${member_index} ${return_success_only}=False ${log_on_success}=True ${log_on_failure}=True ${stderr_must_be_empty}=True
+ [Documentation] Obtain Ip address, open session, call SSHKeywords.Execute_Command_Passes, close session and return output. This affects which SSH session is active.
+ ${member_ip} = Resolve_Ip_Address_For_Member ${member_index}
+ BuiltIn.Run_Keyword_And_Return SSHKeywords.Run_Unsafely_Keyword_Over_Temporary_Odl_Session ${member_ip} Execute_Command_Passes ${command} return_success_only=${return_success_only} log_on_success=${log_on_success}
+ ... log_on_failure=${log_on_failure} stderr_must_be_empty=${stderr_must_be_empty}
+
Run_Bash_Command_On_List_Or_All
[Arguments] ${command} ${member_index_list}=${EMPTY}
[Documentation] Cycle through indices (or all), run command on each.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ # TODO: Migrate callers to Check_Bash_Command_*
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list}
\ Run_Bash_Command_On_Member command=${command} member_index=${index}
Run_Bash_Command_On_Member
[Arguments] ${command} ${member_index}
- [Documentation] Obtain IP, call Utils and return output. This does not preserve active ssh session.
- # TODO: Rename these keyword to Run_Bash_Command_On_Member to distinguish from Karaf (or even Windows) commands.
+ [Documentation] Obtain IP, call Utils and return output. This keeps previous ssh session active.
+ # TODO: Migrate callers to Check_Bash_Command_*
${member_ip} = Collections.Get_From_Dictionary dictionary=${ClusterManagement__index_to_ip_mapping} key=${member_index}
${output} = SSHKeywords.Run_Keyword_Preserve_Connection Utils.Run_Command_On_Controller ${member_ip} ${command}
+ Log ${output}
[Return] ${output}
Run_Karaf_Command_On_List_Or_All
[Arguments] ${command} ${member_index_list}=${EMPTY} ${timeout}=10s
[Documentation] Cycle through indices (or all), run karaf command on each.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list}
\ ${member_ip} = Collections.Get_From_Dictionary dictionary=${ClusterManagement__index_to_ip_mapping} key=${index}
\ KarafKeywords.Safe_Issue_Command_On_Karaf_Console ${command} ${member_ip} timeout=${timeout}
Install_Feature_On_List_Or_All
[Arguments] ${feature_name} ${member_index_list}=${EMPTY} ${timeout}=60s
[Documentation] Attempt installation on each member from list (or all). Then look for failures.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
${status_list} = BuiltIn.Create_List
: FOR ${index} IN @{index_list}
\ ${status} ${text} = BuiltIn.Run_Keyword_And_Ignore_Error Install_Feature_On_Member feature_name=${feature_name} member_index=${index}
With_Ssh_To_List_Or_All_Run_Keyword
[Arguments] ${member_index_list} ${keyword_name} @{args} &{kwargs}
[Documentation] For each index in given list (or all): activate SSH connection, run given Keyword, close active connection. Return None.
- ... Note that if the Keyword affects SSH connections, results are still deterministic, but perhaps undesirable.
... Beware that in order to avoid "got positional argument after named arguments", first two arguments in the call should not be named.
BuiltIn.Comment This keyword is experimental and there is high risk of being replaced by another approach.
# TODO: For_Index_From_List_Or_All_Run_Keyword applied to With_Ssh_To_Member_Run_Keyword?
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ # TODO: Imagine another keyword, using ScalarClosures and adding member index as first argument for each call. Worth it?
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${member_index} IN @{index_list}
\ ${member_ip} = Resolve_IP_Address_For_Member ${member_index}
- \ SSHKeywords.Open_Connection_To_Odl_System ip_address=${member_ip}
- \ BuiltIn.Run_Keyword ${keyword_name} @{args} &{kwargs}
- \ SSHLibrary.Close_Connection
+ \ SSHKeywords.Run_Unsafely_Keyword_Over_Temporary_Odl_Session ${member_ip} ${keyword_name} @{args} &{kwargs}
Safe_With_Ssh_To_List_Or_All_Run_Keyword
[Arguments] ${member_index_list} ${keyword_name} @{args} &{kwargs}
SSHKeywords.Run_Keyword_Preserve_Connection With_Ssh_To_List_Or_All_Run_Keyword ${member_index_list} ${keyword_name} @{args} &{kwargs}
Clean_Directories_On_List_Or_All
- [Arguments] ${member_index_list}=${EMPTY} ${directory_list}=${EMPTY} ${karaf_home}=${KARAF_HOME}
+ [Arguments] ${member_index_list}=${EMPTY} ${directory_list}=${EMPTY} ${karaf_home}=${KARAF_HOME} ${tmp_dir}=${EMPTY}
[Documentation] Clear @{directory_list} or @{ODL_DEFAULT_DATA_PATHS} for members in given list or all. Return None.
+ ... If \${tmp_dir} is nonempty, use that location to preserve data/log/.
... This is intended to return Karaf (offline) to the state it was upon the first boot.
${path_list} = Builtin.Set Variable If "${directory_list}" == "${EMPTY}" ${ODL_DEFAULT_DATA_PATHS} ${directory_list}
+ BuiltIn.Run_Keyword_If """${tmp_dir}""" != "" Check_Bash_Command_On_List_Or_All mkdir -p '${tmp_dir}' && rm -vrf '${tmp_dir}/log' && mv -vf '${karaf_home}/data/log' '${tmp_dir}/' ${member_index_list}
Safe_With_Ssh_To_List_Or_All_Run_Keyword ${member_index_list} ClusterManagement__Clean_Directories ${path_list} ${karaf_home}
+ BuiltIn.Run_Keyword_If """${tmp_dir}""" != "" Check_Bash_Command_On_List_Or_All mkdir -p '${karaf_home}/data' && rm -vrf '${karaf_home}/log' && mv -vf '${tmp_dir}/log' '${karaf_home}/data/' ${member_index_list}
+
+Store_Karaf_Log_On_List_Or_All
+ [Arguments] ${member_index_list}=${EMPTY} ${dst_dir}=/tmp ${karaf_home}=${KARAF_HOME}
+ [Documentation] Saves karaf.log to the ${dst_dir} for members in given list or all. Return None.
+ Safe_With_Ssh_To_List_Or_All_Run_Keyword ${member_index_list} SSHKeywords.Execute_Command_Should_Pass cp ${karaf_home}/data/log/karaf.log ${dst_dir}
+
+Restore_Karaf_Log_On_List_Or_All
+ [Arguments] ${member_index_list}=${EMPTY} ${src_dir}=/tmp ${karaf_home}=${KARAF_HOME}
+ [Documentation] Places stored karaf.log to the ${karaf_home}/data/log for members in given list or all. Return None.
+ Safe_With_Ssh_To_List_Or_All_Run_Keyword ${member_index_list} SSHKeywords.Execute_Command_Should_Pass cp ${src_dir}/karaf.log ${karaf_home}/data/log/
ClusterManagement__Clean_Directories
[Arguments] ${relative_path_list} ${karaf_home}
[Arguments] ${uri} ${expected_data} ${member_index_list}=${EMPTY}
[Documentation] Send a GET with the supplied uri to all or some members defined in ${member_index_list}.
... Then check received data is = ${expected data}.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list}
\ ${data} = Get_From_Member uri=${uri} member_index=${index}
\ TemplatedRequests.Normalize_Jsons_And_Compare ${expected_data} ${data}
[Arguments] ${uri} ${dictionary} ${member_index_list}=${EMPTY}
[Documentation] Send a GET with the supplied uri to all or some members defined in ${member_index_list}.
... Then check received for occurrences of items expressed in a dictionary ${dictionary}.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list}
\ ${data} = Get_From_Member uri=${uri} member_index=${index}
\ Utils.Check Item Occurrence ${data} ${dictionary}
[Arguments] ${uri} ${member_index_list}=${EMPTY}
[Documentation] Send a GET with the supplied uri to all or some members defined in ${member_index_list}.
... Then check there is no content.
- ${index_list} = ClusterManagement__Given_Or_Internal_Index_List given_list=${member_index_list}
+ ${index_list} = List_Indices_Or_All given_list=${member_index_list}
: FOR ${index} IN @{index_list}
\ ${session} = Resolve_Http_Session_For_Member member_index=${index}
\ Utils.No_Content_From_URI ${session} ${uri}
${ip_address} = Collections.Get From Dictionary dictionary=${ClusterManagement__index_to_ip_mapping} key=${member_index}
[Return] ${ip_address}
+Resolve_IP_Address_For_Members
+ [Arguments] ${member_index_list}
+ [Documentation] Return a list of IP address of given indexes.
+ ${member_ip_list} = BuiltIn.Create_List
+ : FOR ${index} IN @{member_index_list}
+ \ ${ip_address} = Collections.Get From Dictionary dictionary=${ClusterManagement__index_to_ip_mapping} key=${index}
+ \ Collections.Append_To_List ${member_ip_list} ${ip_address}
+ [Return] ${member_ip_list}
+
Resolve_Http_Session_For_Member
[Arguments] ${member_index}
[Documentation] Return RequestsLibrary session alias pointing to node of given index.
${sync_status} = Collections.Get_From_Dictionary dictionary=${value_object} key=SyncStatus
[Return] ${sync_status}
-List_Indices_Minus_Member
- [Arguments] ${member_index} ${member_index_list}=${EMPTY}
- [Documentation] Create a new list which contains indices from ${member_index_list} (or all) without ${member_index}.
- ${index_list} = ClusterManagement__Given_Or_Empty_List ${member_index_list}
- Collections.Remove Values From List ${index_list} ${member_index}
- [Return] ${index_list}
+List_All_Indices
+ [Documentation] Create a new list of all indices.
+ BuiltIn.Run_Keyword_And_Return List_Indices_Or_All
-ClusterManagement__Given_Or_Internal_Index_List
+List_Indices_Or_All
[Arguments] ${given_list}=${EMPTY}
[Documentation] Utility to allow \${EMPTY} as default argument value, as the internal list is computed at runtime.
- ... This keyword always return a (shallow) copy of given or default list,
+ ... This keyword always returns a (shallow) copy of given or default list,
... so operations with the returned list should not affect other lists.
... Also note that this keyword does not consider empty list to be \${EMPTY}.
- ... TODO: This keyword is frequently used for obtaining copy of ${ClusterManagement__member_index_list}. Give this keyword public name.
${return_list_reference} = BuiltIn.Set_Variable_If """${given_list}""" != "" ${given_list} ${ClusterManagement__member_index_list}
${return_list_copy} = BuiltIn.Create_List @{return_list_reference}
[Return] ${return_list_copy}
-ClusterManagement__Given_Or_Empty_List
- [Arguments] ${given_list}=${EMPTY}
- [Documentation] Utility to allow \${EMPTY} as default argument value, as an empty list is computed at runtime.
- ${empty_list} = BuiltIn.Create_List
- ${given_length} = BuiltIn.Get_Length ${given_list}
- ${return_list} = BuiltIn.Set_Variable_If ${given_length} > 0 ${given_list} ${empty_list}
- [Return] ${return_list}
+List_Indices_Minus_Member
+ [Arguments] ${member_index} ${member_index_list}=${EMPTY}
+ [Documentation] Create a new list which contains indices from ${member_index_list} (or all) without ${member_index}.
+ ${index_list} = List_Indices_Or_All ${member_index_list}
+ Collections.Remove Values From List ${index_list} ${member_index}
+ [Return] ${index_list}
ClusterManagement__Compute_Derived_Variables
- [Arguments] ${int_of_members}
+ [Arguments] ${int_of_members} ${http_timeout}=${DEFAULT_TIMEOUT_HTTP} ${http_retries}=0
[Documentation] Construct index list, session list and IP mapping, publish them as suite variables.
@{member_index_list} = BuiltIn.Create_List
@{session_list} = BuiltIn.Create_List
&{index_to_ip_mapping} = BuiltIn.Create_Dictionary
: FOR ${index} IN RANGE 1 ${int_of_members+1}
- \ ClusterManagement__Include_Member_Index ${index} ${member_index_list} ${session_list} ${index_to_ip_mapping}
+ \ ClusterManagement__Include_Member_Index ${index} ${member_index_list} ${session_list} ${index_to_ip_mapping} http_timeout=${http_timeout}
+ \ ... http_retries=${http_retries}
BuiltIn.Set_Suite_Variable \${ClusterManagement__member_index_list} ${member_index_list}
BuiltIn.Set_Suite_Variable \${ClusterManagement__index_to_ip_mapping} ${index_to_ip_mapping}
BuiltIn.Set_Suite_Variable \${ClusterManagement__session_list} ${session_list}
ClusterManagement__Include_Member_Index
- [Arguments] ${index} ${member_index_list} ${session_list} ${index_to_ip_mapping}
+ [Arguments] ${index} ${member_index_list} ${session_list} ${index_to_ip_mapping} ${http_timeout}=${DEFAULT_TIMEOUT_HTTP} ${http_retries}=0
[Documentation] Add a corresponding item based on index into the last three arguments.
... Create the Http session whose alias is added to list.
Collections.Append_To_List ${member_index_list} ${index}
Collections.Set_To_Dictionary ${index_to_ip_mapping} ${index} ${member_ip}
# Http session, with ${AUTH}, without headers.
${session_alias} = Resolve_Http_Session_For_Member member_index=${index}
- RequestsLibrary.Create_Session ${session_alias} http://${member_ip}:${RESTCONFPORT} auth=${AUTH} max_retries=0
+ RequestsLibrary.Create_Session ${session_alias} http://${member_ip}:${RESTCONFPORT} auth=${AUTH} timeout=${http_timeout} max_retries=${http_retries}
Collections.Append_To_List ${session_list} ${session_alias}
+
+Sync_Status_Should_Be_False
+ [Arguments] ${controller_index}
+ [Documentation] Verify that cluster node is not in sync with others
+ ${status} Get_Sync_Status_Of_Member ${controller_index}
+ BuiltIn.Should_Not_Be_True ${status}
+
+Sync_Status_Should_Be_True
+ [Arguments] ${controller_index}
+ [Documentation] Verify that cluster node is in sync with others
+ ${status} Get_Sync_Status_Of_Member ${controller_index}
+ BuiltIn.Should_Be_True ${status}
+
+Return_Member_IP
+ [Arguments] ${member_index}
+ [Documentation] Return the IP address of the member given the member_index.
+ ${member_int} = BuiltIn.Convert_To_Integer ${member_index}
+ ${member_ip} = Collections.Get_From_Dictionary dictionary=${ClusterManagement__index_to_ip_mapping} key=${member_int}
+ [Return] ${member_ip}
+
+Check Service Status
+ [Arguments] ${odl_ip} ${system_ready_state} ${service_state} @{service_list}
+ [Documentation] Issues the karaf shell command showSvcStatus to verify the ready and service states are the same as the arguments passed
+ ${service_status_output} = BuiltIn.Run Keyword If ${NUM_ODL_SYSTEM} > 1 KarafKeywords.Issue_Command_On_Karaf_Console showSvcStatus -n ${odl_ip} ${odl_ip} ${KARAF_SHELL_PORT}
+ ... ELSE KarafKeywords.Issue_Command_On_Karaf_Console showSvcStatus ${odl_ip} ${KARAF_SHELL_PORT}
+ BuiltIn.Should Contain ${service_status_output} ${system_ready_state}
+ : FOR ${service} IN @{service_list}
+ \ BuiltIn.Should Match Regexp ${service_status_output} ${service} +: ${service_state}
+
+Check Status Of Services Is OPERATIONAL
+ [Arguments] @{service_list}
+ [Documentation] This keyword will verify whether all the services are operational in all the ODL nodes
+ : FOR ${i} IN RANGE ${NUM_ODL_SYSTEM}
+ \ ClusterManagement.Check Service Status ${ODL_SYSTEM_${i+1}_IP} ACTIVE OPERATIONAL @{service_list}
Code Review / integration / test.git / blobdiff