Release notes

Known problems at eGA

• Solution-level problems
• Operations Management on container platforms
• Device Dashboard
• Event Analytics
• Event Search
• Globalization
• Network Health Dashboard
• Operations Analytics - Log Analysis

Solution-level problems

Non-alphanumeric characters do not work in Event Viewer search

If non-alphanumeric characters are used in an event search from the Event Viewer, then these characters are replaced with their ASCII equivalent. This causes the search to be performed with the incorrect characters, and events that should have matched the search term are not returned in the search results. To resolve this, edit the search field and replace the string that has ASCII characters in it with the original search term that you entered.

Event Viewer count is not correct

This issue concerns WebGUI and the Event Viewer. The filter count numbers represented by the colored icons on the toolbar of the Event Viewer, should show the total number of events, including all parent and child events in the list. However, when the user selects a filter, child events that match the filter selected but whose parents events do not, are showing in the count number but not in the list.

Fail to connect to the primary cluster in Netcool/Impact GUI after failback

When an Impact failback happens, the Netcool/Impact GUI is not able to identify which is the active primary cluster. Therefore, when the user logs into the Netcool/Impact GUI, an error will appear. This issue only concerns the user interface and doesn't affect the backend. To solve the problem, the user needs to delete all Netcool/Impact server pods available by using the example command:

kubectl delete pod <release>-nciserver-0 <release>-nciserver-1 ... <release>-nciserver-n

This command synchronize again the system and allow downtime when the Netcool/Impact are being restarted. Once Netcool/Impact is running again, Netcool/Impact is available.

Web GUI on premises is missing menu

This issue occurs when you install Netcool Operations Insight on premises with a single Object Server. After updating Web GUI with Network Manager the navigation panel menu in Device Dashboard is missing. To solve this issue, ensure that Network Manager is updated to the latest fix pack and that Jazz® for Service Management, Device Dashboard and WebSphere® Application Server are running the latest versions.

Unable to create new users in LDAP using WebSphere Application Server

When a new user is created using the WebSphere Application Server, the UniqueName attribute references the defaultFileBasedRealm instead of LDAP. This means that the new user cannot be assigned to groups and therefore cannot be assigned roles in LDAP.

Certificate Verification popup

When logging in to DASH, an Information pop-up is displayed, requesting that certificates for some links be verified. These links do not work. This pop-up can be safely closed.

Operations Management on Container Platforms

Cross-Origin Resource Sharing exception occurs when launchingAgile Service Manager topology viewer

This issue occurs on a hybrid install with Netcool Operations Insight and Agile Service Manager on premises, and Cloud Native Analyticson the cloud. The user is unable to launch to the topology viewer from the incident viewer as a Cross-Origin Resource Sharing exception verifies. The workaround to this issue is to perform the topology launch from the event viewer instead, by finding the corresponding event to the incident, and by clicking the topology icon shown there instead of what is in the incident viewer.

ea-asm-normalizer restarts due to memory leak

This issue is due to a memory leak. The memory size of the ea-asm-normalizerpod is gradually growing, and eventually the pod dies, and restarts. However, the service will not be impacted.

Spark line grows off the page for groups with large numbers of events

This is an issue that occurs when Netcool Operations Insight is deployed at size 0 on a Red Hat OpenShift cluster and trained with a large number of events. In the temporal group viewer, groups with large numbers of events are causing issues with the spark-line which grows off the page preventing the group from being expanded.

Manage Policies Widget

This is an issue that occurs when Netcool Operations Insight is deployed at size 0 on a Red Hat OpenShift cluster and trained with a large number of events. The temporal policies list doesn't provide access to policy inspection and therefore to the timeline graph, membership events or to the approve or reject option. The policies names are not clickable. To avoid this issue it is necessary to open the Event viewer before opening the policy list.

noi-proxy unable to recover after OpenShift upgrade or master nodes reboot

This problem occurs when the user reboots all master nodes or performs an OpenShift upgrade. The noi-proxy looses connection to the Kubernetes API and the proxy pod remains in a broken state. This will prevent remote Netcool/OMNIbus clients to connect. To solve this issue the pod needs to be manually restarted.

For demo/trial deployments, PodDisruptionBudgets (PDBs) block the upgrade of Red Hat OpenShift

When Netcool Operations Insight is deployed at size 0 (demo or trial deployment), PodDisruptionBudgets block the upgrade of OpenShift. PDBs can be deleted by running the following commands before upgrading OpenShift:

oc get pdb | grep <helm-release>
oc delete pdb -l release=noi-release-name --all

The oc get pdb command returns the PDB names, for example: noi-cassandra-pdb. Each PDB must be deleted with the oc delete pdb command.

Netcool Operations Insight pods detect Agile Service Manager SecurityContextConstraints (SCC)

This is a known issue that occurs when you deploy Agile Service Manager and Netcool Operations Insight on a clean Red Hat OpenShift cluster. The configured SCC across the Netcool Operations Insight pods uses the Agile Service Manager SCC, ibm-netcool-asm-prod-scc, and not the intended Netcool Operations Insight SCC. Both Agile Service Manager and Netcool Operations Insight are deployed into the same namespace, with Agile Service Manager deployed before Netcool Operations Insight. Agile Service Manager is deployed across to the default service account with the Agile Service Manager SCC applied across the entire namespace. Therefore, the Pod Admission Controller selects the Agile Service Manager SCC instead of the Netcool Operations Insight one.

Cloud Native Analytics kafka clients fail to re-establish a connection to kafka brokers on kafka restart

When the kafka pod is about to restart, various kafka clients applications fail to re-establish a connection to kafka. These applications presenting this issue are for example inferenceservice, ea-asm-normalizer-normalizerstreams and ea-asm-normalizer-mirrormaker. This is a known issue with the common-kafka containers and occurs when using a size0 deployment. The details of the issue are the following:

WARN  [2020-02-05 10:26:52,183] org.apache.kafka.clients.NetworkClient: [Producer clientId=producer-1] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:52,183] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-2, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:52,183] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-5, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:52,183] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-3, groupId=0] Connection to node 0 could not be established. Broker may not be available.
INFO  [2020-02-05 10:26:54,906] com.ibm.itsm.topology.kafka.client.api.ConsumerWorker: Consumer topic [ea-asm-enriched-events] Read 1 records from Kafka
INFO  [2020-02-05 10:26:54,907] com.ibm.itsm.topology.kafka.client.api.ConsumerWorker: Consumer topic [ea-asm-enriched-events] Read 1 records from Kafka
INFO  [2020-02-05 10:26:54,907] com.ibm.itsm.topology.kafka.client.api.ConsumerWorker: Consumer topic [ea-asm-enriched-events] Read 2 records from Kafka
INFO  [2020-02-05 10:26:54,914] com.ibm.itsm.topology.kafka.client.api.ConsumerWorker: Consumer topic [ea-asm-enriched-events] Read 3 records from Kafka
INFO  [2020-02-05 10:26:54,914] com.ibm.itsm.topology.kafka.client.api.ConsumerWorker: Consumer topic [ea-asm-enriched-events] Read 1 records from Kafka
WARN  [2020-02-05 10:26:55,255] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-2, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:55,255] org.apache.kafka.clients.NetworkClient: [Producer clientId=producer-1] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:55,255] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-5, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:55,256] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-3, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:58,327] org.apache.kafka.clients.NetworkClient: [Producer clientId=producer-1] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:58,327] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-3, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:58,327] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-5, groupId=0] Connection to node 0 could not be established. Broker may not be available.
WARN  [2020-02-05 10:26:58,327] org.apache.kafka.clients.NetworkClient: [Consumer clientId=consumer-2, groupId=0] Connection to node 0 could not be established. Broker may not be available.

To solve the issue, delete the pods that are not working.

Cloud Native Analytics automatic scheduled training doesn't appear to run

This is a known issue. After the installation of the Netcool Operations Insight helm chart and the execution of the runTraining.sh script, Cloud Native Analytics should automatically start training. An error in the analytics configuration files prevents Cloud Native Analytics from producing policies for temporal correlation or seasonality. To solve this issue, run the following command after you installed the helm chart and the pods are successfully running:

RELEASE=helm_release_name;\
TENANT_ID=tenant_id;
timeout 10 kubectl port-forward service/$RELEASE'-ibm-hdm-analytics-dev-trainer' 8080:8080 > /dev/null & \
sleep 2; \
  curl -X DELETE --header "X-TenantID: $TENANT_ID" 127.0.0.1:8080/1.0/training/analytics/related-events/schedule; \
  curl -X DELETE --header "X-TenantID: $TENANT_ID" 127.0.0.1:8080/1.0/training/analytics/seasonal-events/schedule; \
  curl -X PATCH --header "Content-Type: application/json" --header "X-TenantID: $TENANT_ID" 127.0.0.1:8080/1.0/training/analytics/related-events/configuration -d '{"properties":{"ea.trainTimeRange":{"period":8035200000}}}'; \
  curl -X PATCH --header "Content-Type: application/json" --header "X-TenantID: $TENANT_ID" 127.0.0.1:8080/1.0/training/analytics/seasonal-events/configuration -d '{"properties":{"ea.trainTimeRange":{"period":8035200000}}}'; \
timeout 10 kubectl port-forward service/$RELEASE'-couchdb' 5984:5984 > /dev/null & \
sleep 2; \
  curl -X PUT '127.0.0.1:5984/schedule/'$TENANT_ID'__0' -d \
  "{\"retrainingIntervalMinutes\":1440,\"enabled\":true,\"tenantId\":\"$TENANT_ID\",\"jobs\":[[\"related-events\"],[\"seasonal-events\"]],\"jobUUIDs\":[\"4c498f8d-0470-465a-95f8-15a5a1fc1692\",\"6dda268f-14ae-47c5-8c4d-64bbd3ef9d3d\"],\"lastUpdateTime\":0}"

Where:

• helm_release_name is the helm release name of your Operations Management on IBM Cloud Private deployment.
• tenant_id is the internal tenant Id used by event analytics.

This command will remove the error in the configuration files.

Monitoring events are not stored in the historical database

This is a known issue. As a workaround, restart the scala pod.

Topology icons not consistently appearing in the event viewer for Agile Service Manager originating events

Using the Agile Service Manager Rest Observer to publish Agile Service Manager topology with status and inspect the resultant events in the Cloud Native Analytics event viewer, we expect all events associated with topology entities to have a topology icon to facilitate launching in context to the topology view.

However, a high percentage of the events don't have a topology icon when they should.

Filter at temporal group does not work for clear events

In a Temporal Group containing events with a variety of severity, when you click the filter button beside the "Example_IBM_ CloudAnalytics" view and you click "Clear", nothing happens. However, if you click on other severities, for example, "severity Major" it works as expected.

Internet Explorer on hybrid deployment: Not able to open CNEA related windows - "Content was blocked because it was not signed by a valid security certificate"

These are the steps as a workaround for Internet Explorer users:

1. On Internet Explorer > Go to Internet options
2. Go to Advanced tab > Uncheck 'Warn about certificate address mismatch'
3. Login to DASH
4. Pop up window - Click on 'Cloud Analytics'. A new window opens. Select the 'More Information' drop down and select 'Go on to the Webpage'. The Certificate error is in the browser URL bar.
5. In the address/URL bar of the main browser window, click on 'Certificate error' > View Certificate
6. Go to Certificate Path tab > Select 'ingress-operator@' > View Certificate
7. Install Certificate > Select 'Place all certificate in the following store' > Browse > Select 'Trusted Root Certificate Authorities' > OK > Next > Finish
8. A new window opens to ask 'Do you want to install this certificate'. Select Yes

Information message about untrusted content after login

After logging in to Operations Management, an information message is displayed:

This console displays content from the below servers that appear to be using untrusted certificates.

Content from untrusted servers cannot be viewed in the console.

Please click each link below to open a new window where you may determine if you would like to accept the certificate and allow the content to be displayed.

When you are done, you may close the new windows and dismiss this warning.

This is a known issue. The pop up window can be closed without following the links.

Replaying historic events on hybrid installation causes an error

On hybrid installations the replaying of historic events to an SSL enabled ObjectServer is not supported. Attempts to do so will result in a write: broken pipe error.

Multiple icons displayed in Event viewer

The Event viewer displays multiple icons for the same correlation type against each event.

Countdown timer on Incident Viewer is not reset when refresh is clicked

When refresh is clicked the countdown timer is not reset to 0 as it should be. The view is still correctly refreshed.

Policy timeouts with error

When selecting a policy from Insights->Manage Policies, the policy will not load and displays the following error:

An error occurred while fetching data from the server. Please make sure you have an active internet connection. More...

You may sometimes hit this issue if you open a policy which has a large number of event instances.

Analytics services generating large log files

Disks can fill up very fast with log files for each service instance. As a workaround, increase disk space in the docker containers space (the /var/lib/docker directory).

Blank window displayed when selecting Event Viewer from side bar of new tab in Internet Explorer V11

When you open the Incident Viewer and select a new tab, the new tab opens displaying the Incident view and a side bar with Event Viewer and Policies. If you select Event Viewer from the side bar, a blank window is displayed.

Default groups are missing from Tools Configuration > Access Criteria

When you log in to IBM Netcool/OMNIbus Web GUI and go to Tools Configuration > Access Criteria, the default groups, such as icpadmins, are missing under the "Access Criteria" section.

Authentication failure - cannot log in to Web GUI when external LDAP is configured

Unable to login to Web GUI using LDAP credentials. Configuring to an external LDAP server is not supported in this release.

Authentication failure - ncobackup cannot authenticate user "root" when ncoprimary is restarted

The following errors are listed in the ncobackup pod logs during startup:

Server 'AGG_B' initialised - entering RUN state.
Updating default root password
2019-06-13T22:11:43: Error: E-OBX-102-023: Failed to authenticate user root. (-3600:Denied)
2019-06-13T22:11:43: Error: E-OBX-102-057: User root@<releasename>-ncobackup-0 failed to login: Denied
Failed to connect
Error: Failed to get login token
Unable to update root password

The following error is seen in the objectserver logfile:

cat /opt/IBM/tivoli/netcool/omnibus/log/AGG_B_audit_file.log
2019-06-13T22:11:43: Error: E-SEC-010-002: authentication failure - cannot authenticate user "root" : Denied

This error can be ignored. The ncobackup pod tries to change the default password and fails, because it has already been changed to the newer version.

UI throws "Pod Security Conflict" warning when selecting namespace during a deployment of Operations Management on IBM Cloud Private with Red Hat OpenShift

The following warnings are displayed:

Pod Security Conflict This chart requires a namespace with a ibm-priviliged-psp pod security policy
Pod Security Warning Your ICP cluster is running all namespaces Unrestricted (ibm-anyuid-hostpath-psp) by default. This could pose a security risk

Red Hat OpenShift does not use pod security policies (PSP), and instead uses security constraint contexts (SCCs). This warning can be ignored, if you have correctly configured a security constraint context in Configuring pod access control.

loadSampleData.sh script fails on IBM Cloud Private with OpenShift

The loadSampleData.sh fails to run on IBM Cloud Private with OpenShift as a result of the aggregation dedup service pod timing out it's redis connection. After a while, the service stops connecting if redis is down. To work around this issue, restart the stateless aggregation dedup pods. Run the kubectl get pod | grep dedup command to get a list of dedup pods, then run the kubectl delete pod command for each dedup pod.

Auto-deploy mode is enabled after running the loadSampleData.sh script.

If you have installed Operations Management on IBM Cloud Private in manual deploy mode, running the loadSampleData.sh script switches your configuration to auto-deploy mode. For more information about manually deploying policies, see the Temporal Group Policies Deploy First parameter in Configuring Installation Parameters.

To disable auto-deploy mode, re-run a portion of the training by running the runTraining.sh script.

Get the start and end times of the sample data as in the following example:

kubectl delete pod ingesnoi3;
kubectl run ingesnoi3 -i --restart=Never --env=LICENSE=accept --image=<repository>/ea/ea-events-tooling: getTimeRange.sh samples/demoTrainingData.json.gz
pod "ingesnoi3" deleted
{"minfirstoccurence":{"epoc":1552023064603,"formatted":"2019-03-08T05:31:04.603Z"},"maxlastoccurrence":{"epoc":1559729860924,"formatted":"2019-06-05T10:17:40.924Z"}}

Re-run the training with seasonality disabled, as in the following example:

kubectl delete pod ingesnoi3;
kubectl run ingesnoi3 -i --restart=Never --env=LICENSE=accept --image-pull-policy=Always --image=<repository>/ea/ea-events-tooling: runTraining.sh -- -r test-install -a SEASONALITY -s 1552023064603 -e 1559729860924 -d false

Where <repository> is the repository name and -s and -e are the start and end times returned by the getTimeRange() command and -d false disables live policy updates.

Authentication errors reported in logs when ncoprimary and ncobackup pods are restarted.

The following errors can be seen in the ncoprimary and ncobackup pod logs when the ncoprimary and ncobackup pods startup:

Information: I-SEC-104-002: Cannot authenticate user "root": Denied
Error: E-OBX-102-023: Failed to authenticate user root. (-3600:Denied)
Error: E-OBX-102-057: User root@dev401-ncoprimary-0 failed to login: Denied
Failed to connect
Error: Failed to get login token
Unable to update root password

These errors are caused by the objectserver starting up and attempting to change the default root password, which has already been changed.

Unable to make TLS connections using administrator GUI versions earlier than V8.1.0.18

Versions of administrator GUI nco-config earlier than V8.1.0.18 fail to make a TLS connection. This is a known issue.

If you want to use the secrets that are automatically created by Operations Management on IBM Cloud Private, you must disable FIPS mode

If the certificate is automatically generated and FIPS mode is specified with the $NCHOME/etc/security/fips.conf file, the Netcool/OMNIbus client fails to make a connection. The CT-LIBRARY error error message is displayed. If you want to use the secrets that are automatically created by Operations Management on IBM Cloud Private then you must disable FIPS mode. Do this by removing the following file: $NCHOME/etc/security/fips.conf file.

Additional Insight Pack installation does not persist

After you install an additional Insight Pack with Operations Analytics - Log Analysis and restart your pod, the Insight Pack disappears. This is a known issue. As a workaround, reinstall the Insight Pack.

Operations Analytics - Log Analysis dashboard does not persist

After you create a Operations Analytics - Log Analysis dashboard with a search result widget and restart your pod, the dashboard disappears. This is a known issue. As a workaround, recreate the dashboard.

Outage of Event Analytics pages or features

Following an Impact Server failover, Event Analytics pages or features might not be available for a short time. This is a known issue.

Cannot change openldap suffix or base distinguished name

When Netcool Operations Insight is running on IBM Cloud Private, and you are using the built-in LDAP server container openldap that is provided with Operations Management on IBM Cloud Private then you must not change the LDAP suffix, which is part if the LDAP base DN, or the LDAP bind distinguished name (bind DN). These are hardcoded based on the values that you specified during installation. If you do try to change these values then the Dashboard Application Services Hub pod, webgui, will not start.

Storage requirements

You must have sufficient storage, as described in Other requirements. When you install Operations Management on IBM Cloud Private, you must specify the Enable Data Persistence option, as described in Configuring Installation Parameters.

Remote access to Db2® container

It is not possible to remotely access the Db2 container (db2ese) that is running within an IBM Cloud Private cluster. You might want to do this if you are using a database client, rather than using the command line within the Db2 container.

Remote access to Netcool/Impact Derby database

It is not possible to remotely access the Netcool/Impact Derby database that is running within an IBM Cloud Private cluster.

New method to access the status information of Netcool/Impact Name Server instances

The method to access the status information of Netcool/Impact Name Server instances has changed from the method documented in the Netcool/Impact documentation at https://www-03preprod.ibm.com/support/knowledgecenter/SSSHYH_7.1.0.16/com.ibm.netcoolimpact.doc/admin/imag_gui_server_viewing_nameserver_status_c.html. If you are running Netcool/Impact under Operations Management on IBM Cloud Private v1.5.0.1, then use the following method to access this status information:

1. Run the following command:

   helm status release-name --tls

   Where release-name is the release name for your Operations Management on IBM Cloud Private deployment.

   This command generates output to your console.

2. In the output navigate to the Impact Servers section, which looks similar to the following snippet:

   Impact Servers:
     Update your hosts file (On the machine you are running your Browser) or your DNS settings with this mapping
     $NODE_IP nci-0.release-name.master-node
     $NODE_IP nci-1.release-name.master-node
   
     firefox https://nci-0.release-name.master-node/nameserver/services
     firefox https://nci-1.release-name.master-node/nameserver/services 

3. Copy and paste the URLs into your browser to retrieve the Netcool/Impact name server instance status.

Device Dashboard

Device Dashboard is unable to differentiate between anomaly thresholds configured for poll definitions that have the same metric name

The Device Dashboard is unable to differentiate between anomaly thresholds configured for poll definitions that have the same metric name. To ensure that metrics are correctly distinguished in the Device Dashboard Performance Insights portlet ensure that any poll definitions created have a unique metric name.

However, this workaround does not apply to the Default Chassis Ping and Default Interface Ping poll definitions, which by default both use a metric with the name PingTime. The consequences of this are best illustrated using an example:

1. Assume you set and enable an anomaly threshold on the Default Chassis Ping poll definition, with the intention of viewing anomalies against the chassis ping time metric.
2. Assume also that you do not set an anomaly threshold on the Default Interface Ping poll definition, as you do not want to view anomalies against the interface ping time metric.
3. However, the threshold you set and enabled on the Default Chassis Ping poll definition will apply to both the chassis and the interface ping time metrics. Consequently in the Interfaces tab of the Performance Insights portlet, selecting PingTime from the Metric drop-down list will return unexpected anomalies in the portlet.

The workaround is to set and enable the same thresholds on both the Default Chassis Ping and Default Interface Ping poll definitions.

Tooltip on sparkline in Performance Insights portlet displays unnecessary scroll bar

The sparklines in the Performance Insights portlet each have an associated tooltip. Some of these tooltips display an unnecessary vertical scrollbar. This issue can safely be ignored.

Performance Insights portlet: sorting the data by Value does not work

In the Interfaces tab within the Performance Insights portlet sorting the data by clicking the Value column does not work.

When launching a new Device Dashboard by right-clicking an event in the Event Viewerwithin an existing Device Dashboard the Device Dashboard does not refresh with the new entity identifier

When launching a new Device Dashboard by right-clicking an event in the Event Viewer within an existing Device Dashboard the newly launched Device Dashboard does not refresh with the entity identifier corresponding to the selected event. To work around this issue, select Event Viewer from the Dashboard Application Services Hub navigation, select the relevant event and launch the Device Dashboard from there.

Integration problem: Widgets do not open when selecting "Show Traffic"

When right-clicking and selecting Show Traffic from any network topology (for example, Network Hop View), widgets do not open. To work around this issue, find the device in Event Viewer , right-click and select Show Traffic.

Device Dashboard does not launch from a Network Hop View opened within a new browser tab

If you launch a Network Hop View within a new browser tab by running the Find in > Network Hop View from any other GUI, then you will not be able to launch the Device Dashboard from that Network Hop View.

Performance Insights portlet filter values are cleared on refresh of Device Dashboard

You can apply filters in the Device Dashboard Performance Insights portlet, in both the Device and Interfaces tabs. However, when the Device Dashboard is refreshed, any values entered into the Interfaces tab filter field are automatically cleared. You will need to enter filter values again after the refresh.

Note: This issue does not affect values entered into the Device tab filter field.

Performance Insights portlet filter values are cleared when switching tabs

When you switch from the Interfaces to the Device tab in the Device Dashboard Performance Insights portlet, any values entered into the Interfaces tab filter field are automatically cleared. You will need to enter filter values again.

Note: This issue does not apply when you switch from the Device to the Interfaces tab.

When portlet preferences are saved Device Dashboard portlets are not displayed properly

After changing portlet preferences in the following Device Dashboard portlets, unexpected results occur in the following portlets:

• Performance Insights portlet: metric and count data disappear from the portlet, and the preferences do not take effect.
• Performance Timeline portlet: timeline data disappears; however, the portlet preferences are saved and are automatically applied when graph is next rendered, either automatically on the next refresh or by clicking a sparkline.

To resolve this issue, do the following:

• Performance Insights portlet: click a device in the Topology portlet. This forces a manual refresh of the Performance Insights portlet.
• Performance Timeline portlet: click any sparkline in the Performance Insights portlet. This reinstates the Performance Timeline portlet and refreshes the content of the portlet.

Occasionally the content of the Device Dashboard Performance Insights portlet does not update when you select a different interface metric

When selecting a new interface metric from the Metric drop-down list in the Performance Insights portlet Interfaces tab, the content of portlet occasionally does not update. To work around this issue, select the metric twice from the drop-down list.

Event Analytics

If your problem is not listed in this section, then refer to Troubleshooting Event Analytics for additional issues.

If one group in a pattern is moved to a new state, the other groups in the pattern disappear from the UI

Symptom:

If one group in a pattern is moved to a new state, the other groups in the pattern disappear from the UI. The pattern and selected group are moved to the new state, but the other groups in the pattern are not visible.

Resolution:

If this situation occurs, to display all the patterns groups, move the group or pattern back to the original state.

If it is necessary to move some groups allocated to a pattern to a new state, you should delete the pattern before the groups are moved. Then, the pattern can be re-created for the same Event Types as before. This method re-allocates the remaining groups that were previously in the pattern.

In the View Related Events portlet the Related Event Groups panel right-click menu commands momentarily appear in capital letters

If you perform the following sequence of steps in the View Related Events portlet then the commands in the Related Event Groups panel right-click menu momentarily appear in full upper-case.

1. Run a configuration.
2. Navigate to the View Related Events portlet.
3. On the Related Event Groups panel, right-click a group and select to either deploy, watch, or archive the group.
4. As soon as Deploy, Watch, or Archive is selected and before the group has moved, right click the related events group again. The right-click menu appears with capital letters.

Unable to add aggregate fields using the Event Analytics Setup Wizard

When running the Event Analytics Setup Wizard, if you add aggregate fields in the Report fields section of the wizard, and save the wizard data, subsequent configuration runs do not display the new fields in seasonal or related event reports. In order to add these fields you must also restart the Netcool/Impact GUI server.

It is not possible to determine which event triggers a pattern from the View Related Events portlet

If you run a configuration and then deploy a pattern, then trigger the pattern so that the trigger statistics are updated, when you subsequently display the View Related Events portlet, it is not possible to determine which event caused the pattern to trigger.

During installation of Event Analytics if the Derby database becomes corrupted reset it to default state

During the installation of Event Analytics, if the Derby database becomes corrupted you must reset it to default state by performing the following steps:

1. Navigate to $IMPACT_HOME/install/dbcore and find the zip archive named ImpactDB_NOI_FP15.zip.
2. Stop of the Netcool/Impact servers in the cluster, including the primary and secondary servers running the Derby database.
3. Back up the existing directory structure in $IMPACT_HOME/db/Server_Name/derby, where Server_Name is the name of the Netcool/Impact server. Once the backup is complete, remove the directory structure.
4. Copy the ImpactDB_NOI_FP15.zip zip file found in step 1 to $IMPACT_HOME/db/Server_Name/derby and unzip tin there.
5. Start the primary Netcool/Impact server and allow it time to fully initialize.
6. Start the secondary Netcool/Impact server and allow it to resynchronize from the primary server.

Ensure that the property impact.featureToStringAsSource.enabled is set to true

The property impact.featureToStringAsSource.enabled must be set to true in order for Event Analytics to work correctly. Set the following value in the Netcool/Impact properties file and run the nci_trigger command to activate the property, as described in the following topic: Netcool/Impact Knowledge Center: nci_trigger.

impact.featureToStringAsSource.enabled=true

Event Analytics Configuration wizard reports error following failover of nciserver-0 pod

When Netcool Operations Insight is running on IBM Cloud Private, and the nciserver-0 container fails over and back again, the Event Analytics Configuration wizard will subsequently report a java.io.FileNotFoundException error. To resolve this issue, regenerate the Netcool/Impact policy list by running the createPolicyList script, as described in the following topic: Netcool/Impact Knowledge Center: createPolicyList.

Seasonality rule generates errors if only Minute of hour is set

If you create a seasonality rule where the only clause set is Minute of hour, then the rule will not fire and will generate errors. An example of this is the following:

minute of hour = 2, 3, 4

To work around this issue, add to the rule a clause that will always be true; for example:

day of week = Mon, Tue, Wed, Thurs, Fri, Sat, Sun
minute of hour = 2, 3, 4

Attempt to edit seasonality rule returns blank screen

When deploying a seasonality rule and then attempting to edit the rule, sometimes the Modify Existing Rule screen opens but is blank, and you are unable to edit the rule.

Order of columns in the Configure Analytics screen is inconsistent

The order of columns in the Configure Analytics screen is inconsistent, and is as follows. The columns are numbered here for ease of reference; they are not numbered in the GUI.

1. Name
2. Event Identity
3. Seasonality Status
4. Related Event Status
5. Start Time
6. End Time
7. Seasonality Phase
8. Seasonality Phase Progress
9. Related Event Phase
10. Related Event Phase Progress
11. Scheduled
12. Related Event Last Run Duration
13. Seasonal Event Last Run Duration

Most of the columns follow the order Seasonality/Seasonal Event, then Related Event. For example, see column pairs 3-4, 7-9, 8-10. However columns 12-13 follow the opposite order (Related Event then Seasonal Event). If you have lots of rows in the table and you scroll down so that the column titles are no longer visible, you might get the Related Event Last Run Duration and Seasonal Event Last Run Duration mixed up. Be aware that the order for these two column is inconsistent with the other columns.

Seasonal event not matching original selection when opting to trigger rule off events related to the seasonal event

When you create a rule based on a seasonal event, if that seasonal event has associated related events then you have the option to trigger the rule based on one or more of those related events, by selecting the Edit Selection... button in the Create Rule screen. When you select this option, the originally selected seasonal event should be in the list and should be selected. Occasionally you might find that the event selected is not the originally selected seasonal event. In this case, you need to search for the desired seasonal event and reselect it.

Related events configuration run errors following adding of aggregate fields

After you have configured the system to add an aggregate field to the analytics reports using the Report fields panel in the Configure Event Analytics wizard screen, you might subsequently encounter errors when you run a related events configuration.

If this is the case then use nci_trigger to export the configuration and check what the Actionable attribute for your new aggregate field is set to.

• If the Actionable attribute is set to "false", then contact IBM Support.
• If the Actionable attribute is set to "true", then set it to "false" and then run nci_trigger to configure the system using the existing settings.

For more information on using nci_trigger to configure the system, see Configuring Event Analytics using the command line.

Error on first use of the Event Analytics Configuration wizard

On first use of the Event Analytics Configuration wizard, you might encounter an error similar to the following:

ATKRST132E An error occurred while transferring a request to the following remote provider: - '404:Cannot access data provider - Impact_NCICLUSTER

To resolve this issue, restart the Netcool/Impact server and Netcool/Impact UI server, then restart the wizard.

Blank screen in Event Analytics Configuration wizard Report fields screen

If, while working through the Event Analytics configuration wizard, you encounter a blank Report fields screen, this might be because you have custom fields in your Historical Event database, and you did one of the following:

• You did not create a database view to map the custom fields to the corresponding standard field names in Netcool/Impact, as described in Mapping customized field names. In this case, perform the following steps:
  1. Exit the wizard.
  2. Create a database view, as described in Mapping customized field names.
  3. Restart the wizard. When you get to the Configure historical event database screen, make sure to select the database view from the History table drop-down list, as described in Configuring the historical event database.
• You created the database view, but when configuring the Historical Event database in the wizard, you specified the unmapped history table instead of the database view.
  1. Click the < Back button in the wizard until you get to the Configure historical event database screen.
  2. In the Configure historical event database screen, make sure to select the database view from the History table drop-down list, as described in Configuring the historical event database.

Unable to export analytics data

When trying to export analytics data, you encounter a "500" error, as follows:

Error 500: java.io.IOException: Unable to validate authorization

If this happens then perform the following steps:

1. Edit the $IMPACT_HOME/etc/<NCI>_server.props file.
2. Set the impact.noi.export.hostname variable to a recognized hostname.

Unable to configure an event pattern in the Event Patterns GUI due to empty Trigger Action section

This occurs if, during the configuration of event pattern processing that is performed as part of Event Analytics system setup, the historical event database column that was chosen for the default event type does not contain any data.

View Related Events > Groups: right-click menu items displayed in capital letters

If you Deploy, Watch, or Archive a related event group from the right-click menu in the View Related Events > Groups panel and then immediately right-click again to access the menu before the group has moved, the menu appears with capital letters. There is no impact on moving groups in this manner.

Time window for Suggested Patterns displays large values

The time window displayed for suggested patterns may occasionally display a very large value in minutes (more than one day in minutes). This is due to anomalies in the data being processed by the event analytics algorithms. If this situation is seen for a suggested pattern, the advised course of action is to either edit the pattern so that the time window is reduced and save this pattern, or to rerun the Event Analytics configuration in the Configure Analytics portlet. This should resolve the issue.

Related events groups and patterns from the Watched tab not moving to the Expired tab

You can modify the expiry time for Watched or Active related events groups. When the expiry time is reached, expired groups and related events from the Watched tab are not moved to the Expired tab within the View Related Events portlet.

Parts of the Events Pattern screen not displaying because an invalid group is selected.

If you attempt to create an event pattern with groups that have EventID as null, the check boxes for Trigger Action, Event Type and Resource Columns do not appear on the Pattern Criteria tab. An error message is displayed if you try to save the pattern. To avoid this issue, ensure that only valid groups are used to create patterns.

View Related Events error: "Failed to load data"

The error "Failed to load data" can appear if you select All from the Configuration pane, when the state is Active or Watched, and there is more than one set of related event configurations with data. To avoid this issue, select the individual configurations instead.

Chart data retrieved from the server is incorrectly formatted

Problems can occur when displaying Seasonal Event Graphs if you are experiencing network latency or a slow network connection between your browser and the DASH/Impact machine. In such a scenario, an incomplete graph page might be displayed with the following error message:

Chart data retrieved from the server is incorrectly formatted
TypeError: Cannot read property 'grammar' of null

Related Events with a medium or weak relationship profile displayed as strong

In the Related Events Details view under More Information, related events with a medium or weak relationship profile might be incorrectly displayed as a strong relationship profile.

Event Analytics configurations are deleted on Netcool/Impact cluster node failover

Following a Netcool/Impact cluster node failover, the running configuration will be correctly stopped but any queued configurations will be lost on the secondary node. These configurations cannot be retrieved and must be recreated on the secondary node.

For more information, see the following technote: http://www.ibm.com/support/docview.wss?uid=swg22012656.

Suggested pattern with a blank Event Type parameter field

When editing a suggested pattern, the Event Type parameter field will be empty if the property type.0.eventtype is set to a value that is empty in the database. To avoid this issue, ensure that the type.0.eventtype property is not set to an empty value in the Event History Database. Selecting an event type field that contains all null values in the history database will result in the pattern criteria section of the create or edit pattern screen appearing blank.

Blank fields during creation of synthetic event on non-occurrence.

Users are given the option to supply one or more values for additional columns by selecting Set additional fields. With the exception of the values specified in the Create Event window, only the values of Node and Summary are copied into the synthetic event.

Removing Netcool/Impact data models that are no longer needed.

This issue refers to the following Netcool/Impact data models that are no longer needed after you upgrade to the IBM Netcool Operations Insight fix pack. Use the Netcool/Impact GUI to remove these data models:

• ObjectServerHistoryDb2ForSeasonality
• ObjectServerHistoryOrclForSeasonality
• ObjectServerHistoryMSSQLForSeasonality

Note: If you choose not to remove these data models, there is no impact to the Event Analytics functionality. You would remove these data models only for cosmetic purposes.

Event summary is truncated

The event summary is occasionally truncated in the Related Events Details portlet Timeline view. To view the event summary, modify the screen resolution temporarily.

Seasonal event rule with time condition does not run

A seasonal event rule, with an action to run after a specific time that is more than 25 seconds, is not run. To ensure that a seasonal event rule that is scheduled to run after a specific time runs correctly, select a time condition of less than or equal to 25 seconds.

Edit selection window hangs when 'selecting all' for a large number of related events during seasonal event rule creation

When creating a seasonal event rule for an event with a large number of related events, you can check the Select all related events check box to associate all the related events with the seasonal event rule. The problem occurs when a large number of related events, on the order of 1000 or more, are selected and Edit selection is clicked. The Edit selection window is displayed but it remains in a loading state.

To avoid this issue, split the report into smaller reports and create rules around reports with fewer related events.

Incorrectly displaying 'No event selected' error message

When creating a pattern for related events and clicking Use Selected Event as Template, if you have not selected an event, the system correctly displays the 'No event selected' error message. However, if you then do select an event and click Use Selected Event as Template again, the error message persists

In this case, you can disregard and close the error message. It will not affect the creation of the pattern.

When testing an event pattern collapsed sections do not display correctly when reopened

Following testing of an event pattern, if you collapse the Groups, Group Instances, and Events sections in the Test tab of the Events Patten GUI using the splitters provided and then you expand the sections again, the data columns in the Groups and Events panels become very narrow and the data cannot be read.

To work around this issue, you can do one of the following:

• Resize the window frame. This causes the columns to resize so that the data becomes visible.
• Manually resize the column widths or refresh the page. Either of these actions causes the columns to be displayed correctly again

Event Search

Unable to access Event Search from the Event Viewer Information right-click command

When Netcool Operations Insight is running on IBM Cloud Private, if you are in the Event Viewer, and you open the Event Properties window by double-clicking an event or by right-clicking an event, then clicking Information > Event Search, then you get a network error indicating that the request to access Operations Analytics - Log Analysis failed. As a workaround, you should access Event Search using an alternative right-click method.

Severity colors in Operations Analytics - Log Analysis charts use random colors for event severity values

Note: This issue is fixed in version 1.4.1.2.

Operations Analytics - Log Analysis charts, such as Event Trend by Severity, Event Storm by AlertGroup, and Severity Distribution, use random colors for chart elements that represent event severity values, and do not use the standard event severity colors used in Web GUI Event Viewer. For example, a pie chart in the Severity Distribution chart portlet might contain a blue pie chart segment representing events of critical severity, even though the standard color for critical severity in the Web GUI Event Viewer is red.

Furthermore, if two or more charts are presented on a dashboard, then it is likely that the color used for a given event severity in one chart portlet on that dashboard page will differ from the same event severity in another chart portlet on that same dashboard page.

Note: Each chart has its own severity color legend. Refer to the severity color legend on each chart to determine which colors are used on that chart to denote different severity values.

This use of random colors that differ from the standard event severity colors used in Web GUI Event Viewer is intentional, and is meant to enable you to distinguish between event severity values on a chart.

HeatMap chart types do not render correctly

Note: This issue is fixed in version 1.4.1.2.

The HeatMap type charts fail to render when using IBM Operations Analytics - Log Analysis V1.3.3, displaying the following error: Parameter category is invalid for Heat Map. The error is caused by a difference in the way charts are defined in Operations Analytics - Log Analysis 1.3.2 and 1.3.3.

The following charts are affected:

• In OMNIbusInsightPack > Last Day > Dynamic Event Dashboard:
  • Hotspots by Node and Alert Group
  • Hotsport by AlertGroup and Severity
• In OMNIbus_Static_Dashboard, opened from the Web GUI  EventSearch tab by right-clicking and selecting Show Event Dashboard by Node:
  • Hotspots by Node and AlertGroup
  • Hotspots by AlertGroup and Severity

To correct this problem, edit the following chart specification files, and search for all instances of category and change them to categories:

• $UNITY_HOME/AppFramework/Apps/OMNIbusInsightPack_v1.3.0.2/Last_Day/OMNIbus_Dynamic_Dashboard.app
• $UNITY_HOME/AppFramework/Apps/OMNIbusInsightPack_V1.3.0.2/OMNIbus_Static_Dashboard.app

Right-click keyword search results hidden by default

This issue affects Event Search prior to Operations Analytics - Log Analysis version 1.3.5. When using the Event Search right-click menu within Web GUI and selecting Show keywords and event count, the search results in Event Search are empty. The problem is caused by the Search Patterns pane being hidden by default in the Operations Analytics - Log Analysis user interface.

To view the Show keywords and event count results, click the Restore Section triangle icon on the left side of the Event Search user interface.

Hotspots by Node, AlertKey not displaying.

If the Hotspots by Node, Alert Group and AlertKey chart fails to display in the Last_Month->Operational Status Dashboard the SOLR heap size might need increasing.

Note: The Hotspots by Node, Alert Group and AlertKey chart is CPU intensive and can be slow to render for systems with a large amount of data.

More information: Search runtime exceptions and IBM SmartCloud® Analytics - Log Analysis Performance and Tuning Guide.

Hover values for the Event Trend by Severity charts do not appear to match the axes.

When hovering over a point on a particular severity the values returned might not appear to match the axes on the chart. It is because the hover values represent that severity only, whereas the values on the axes are cumulative. For example, if there are 20 Intermediate severity events and 26 Major severity events displayed on the line above, the Major events will appear against 46 on the Y-axis.

Drill down does not return results.

The drill down function is not available for the type Omnibus Tree Map. This affects some of the charts in the Operational Efficiency dashboard in the Last Month folder, and in the OMNIbus Operations Manager and OMNIbus Spot Important Alerts dashboards in the Last Hour folder.

If drill down is required for these charts you can use the default Tree Map specification instead. To change specification, click the chart settings icon on the right of the chart and change Chart Type from OMNIbus Tree Map to Tree Map.

Help text not rendered correctly for Event Analysis and Reduction in bi-directional locales.

This affects the help text in the Event Analysis and Reduction dashboards for Analyze and Reduce Event Volumes, Introduction to the Apps, and Helpful Links. The help text is not rendered correctly in the Arabic and Hebrew locales.

You can view text directly as an HTML file in a browser that supports bi-directional locales. The relevant files are Analyze_and_reduce_event_volumes.html, Introduction_to_the_Apps.html,and Helpful_links.html. The files are located in $UNITY_HOME/AppFramework/Apps/OMNIbusInsightPack_v1.3.0.2/locale/<LOCALE>/LC_MESSAGES.

Globalization

The following issues affect the non-English language versions of Netcool Operations Insight.

Event Search: count and time stamp hover help are not rendered into a non-English language in some charts

This issue affects the OMNIbusInsightPack_v1.3.1. The following texts are not rendered into a non-English language and appear in English only:

• Hover help for time stamp and count in stacked bar charts, heat maps and pie charts
• Legend for the Count field in the Hotspots by Node and AlertGroup chart and Hotspots by AlertGroup and Severity chart of the xxxOMNIbus Static Dashboard custom app
• Legend for the Count field in the Last Day - Hotspots by AlertGroup and Severity chart, 'Last Day - Event Storm by Node chart, and Last Day - Hotspots by Node and AlertGroup of the OMNIbus Dynamic Dashboard custom app

Network Health Dashboard: In the Top Performers widget, the Japanese rendering of the time to refresh is not displayed correctly

In the Top Performers widget, the Japanese rendering of the time to refresh is not displayed correctly. The correct word order in Japanese has the number preceding the text; however, the word order displayed has the number following the text.

Topology Search: error message only partially rendered into non-English languages

This issue affects the Network Manager Insight Pack V1.3.0.0. If you attempt to run a topology search between two nodes on the same device, the error message that is displayed is only partially rendered into non-English languages. The error message in full is as follows:

An error occurred calculating the path between 'X' and 'X',

The source and destination Node's cannot be the same

(Where X is the value of the NmosObjInst for the device. The first half of the message, An error occurred calculating the path between 'X' and 'X', is rendered into non-English languages. The second half of the message The source and destination Node's cannot be the same is not rendered into non-English languages and always appears in English.

IBM Installation Manager

Console mode: cannot install Netcool/OMNIbus core and Web GUI or Netcool/Impact at the same time.

When you install Netcool/OMNIbus core and Web GUI or Netcool/Impact at the same time with Installation Manager - console mode, the installation paths for Web GUI and Netcool/Impact are not prompted for and the installation fails. If you are performing the installation with Installation Manager - console mode, you must install the components separately.

Note: Installing Jazz for Service Management and IBM WebSphere Application Server is not supported for Installation Manager - console mode.

Network Health Dashboard

The known problems vary depending on which version of Network Manager is installed.

If you have Network Manager 4.2.0.4 installed

There are no known problems for the Network Health Dashboard.

If you have Network Manager 4.2.0.3 installed

The Network Health Dashboard has the following known problems.

Must install Network Health Dashboard with the Network Manager GUI Components

For Fix Packs 1, 2, and 3, if you want to install the Network Health Dashboard, you must install it in the same IBM Installation Manager session as the Network Manager GUI Components. Otherwise, some interaction between widgets and pages might not work.

If you upgrade the Network Manager GUI Components from Fix Pack 1 to 2, or 2 to 3, you must upgrade the Network Health Dashboard in the same IBM Installation Manager session.

If you roll back the Network Manager GUI Components from Fix Pack 2 to Fix Pack 1, you must roll back the Network Health Dashboard in the same IBM Installation Manager session.

If you roll back the Network Manager GUI Components from Fix Pack 3 to Fix Pack 2, you do not need to roll back the Network Health Dashboard in the same IBM Installation Manager session.

If you encounter problems with upgrading or rolling back the Network Manager GUI Components, apply the following workaround: Uninstall the Network Health Dashboard, retry the GUI upgrade or rollback, then reinstall the Network Health Dashboard.

Network Health Dashboard does not work after rolling back from Fix Pack 3

If you update from Network Manager 4.2 to 4.2 Fix Pack 3 and then roll back to 4.2, the Network Health Dashboard does not function. Take a backup of your system before upgrading, and restore the backup instead of rolling back the product.

You must install the Network Health Dashboard with the Network Manager GUI Components

If you want to install the Network Health Dashboard. you must install it in the same IBM Installation Manager session as the Network Manager GUI Components. Otherwise, some interaction between widgets and pages might not work.

You must roll back the Network Manager GUI Components and the Network Health Dashboard packages together

If you have installed the Network Manager 4.2 GA release and the Network Health Dashboard, and you then updated to 4.2 Fix Pack 1, and at this point you want to roll back one or more packages, then you must roll back the Network Manager GUI Components and the Network Health Dashboard packages together. If you roll back one of these packages without the other, there is a risk that your system will become corrupted.

Operations Analytics - Log Analysis

Operations Analytics - Log Analysis installation fails

Operations Analytics - Log Analysis V1.3.x requires a minimum 5 GB disk space on all supported operating systems. If less than 5 GB is found, the installer hangs. No error message is displayed. If this problem occurs, terminate the installer.