立即与支持人员聊天
与支持团队交流

One Identity Safeguard for Privileged Sessions 6.0.2 - Administration Guide

Preface Introduction The concepts of One Identity Safeguard for Privileged Sessions (SPS) The Welcome Wizard and the first login Basic settings
Supported web browsers and operating systems The structure of the web interface Network settings Configuring date and time System logging, SNMP and e-mail alerts Configuring system monitoring on SPS Data and configuration backups Archiving and cleanup Forwarding data to third-party systems Joining to One Identity Starling
User management and access control Managing One Identity Safeguard for Privileged Sessions (SPS)
Controlling One Identity Safeguard for Privileged Sessions (SPS): reboot, shutdown Managing Safeguard for Privileged Sessions (SPS) clusters Managing a high availability One Identity Safeguard for Privileged Sessions (SPS) cluster Upgrading One Identity Safeguard for Privileged Sessions (SPS) Managing the One Identity Safeguard for Privileged Sessions (SPS) license Accessing the One Identity Safeguard for Privileged Sessions (SPS) console Sealed mode Out-of-band management of One Identity Safeguard for Privileged Sessions (SPS) Managing the certificates used on One Identity Safeguard for Privileged Sessions (SPS)
General connection settings HTTP-specific settings ICA-specific settings RDP-specific settings SSH-specific settings Telnet-specific settings VMware Horizon View connections VNC-specific settings Indexing audit trails Using the Search interface Searching session data on a central node in a cluster Advanced authentication and authorization techniques Reports The One Identity Safeguard for Privileged Sessions (SPS) RPC API The One Identity Safeguard for Privileged Sessions (SPS) REST API One Identity Safeguard for Privileged Sessions (SPS) scenarios Troubleshooting One Identity Safeguard for Privileged Sessions (SPS) Using SPS with SPP Configuring external devices Using SCP with agent-forwarding Security checklist for configuring One Identity Safeguard for Privileged Sessions (SPS) Jumplists for in-product help LDAP user and group resolution in SPS Appendix: Deprecated features

Displaying custom connection statistics

The following describes how to display statistics of a specific connection policy.

To display statistics of a specific connection policy

  1. Navigate to Basic Settings > Dashboard > Connection statistics.

  2. To display the statistics of a connection policy, enter the name of the policy into the Connection.

  3. Select the time period to display from the Select resolution field.

  4. Click View graph.

Troubleshooting a One Identity Safeguard for Privileged Sessions (SPS) cluster

The following sections help you to solve problems related to high availability clusters.

Understanding One Identity Safeguard for Privileged Sessions (SPS) cluster statuses

This section explains the possible statuses of the One Identity Safeguard for Privileged Sessions (SPS) cluster and its nodes, the DRBD data storage system, and the heartbeat interfaces (if configured). SPS displays this information on the Basic Settings > High Availability page.

The Status field indicates whether the SPS nodes recognize each other properly and whether those are configured to operate in high availability mode. The status of the individual SPS nodes is indicated in the Node HA state field of the each node. The following statuses can occur:

  • Standalone: There is only one SPS unit running in standalone mode, or the units have not been converted to a cluster (the Node HA state of both nodes is standalone). Click Convert to Cluster to enable High Availability mode.

  • HA: The two SPS nodes are running in High Availability mode. Node HA state is HA on both nodes, and the Node HA UUID is the same on both nodes.

  • Half: High Availability mode is not configured properly, one node is in standalone, the other one in HA mode. Connect to the node in HA mode, and click Join HA to enable High Availability mode.

  • Broken: The two SPS nodes are running in High Availability mode. Node HA state is HA on both nodes, but the Node HA UUID is different. For assistance, contact our Support Team.

  • Degraded: SPS was running in high availability mode, but one of the nodes has disappeared (for example broken down, or removed from the network). Power on, reconnect, or repair the missing node.

  • Degraded (Disk Failure): A hard disk of the secondary node is not functioning properly and must be replaced. To request a replacement hard disk and for details on replacing the hard disk, contact our Support Team.

  • Degraded Sync: Two SPS units were joined to High Availability mode, and the first-time synchronization of the disks is currently in progress. Wait for the synchronization to complete. Note that in case of large disks with lots of stored data, synchronizing the disks can take several hours.

  • Split brain: The two nodes lost the connection to each other, with the possibility of both nodes being active nodes (that is, primary nodes) for a time.

    Caution:

    Hazard of data loss In this case, valuable audit trails might be available on both SPS nodes, so special care must be taken to avoid data loss. For details on solving this problem, see Recovering from a split brain situation.

    Do NOT reboot or shut down the nodes.

  • Invalidated: The data on one of the nodes is considered out-of-sync and should be updated with data from the other node. This state usually occurs during the recovery of a split-brain situation when the DRBD is manually invalidated.

  • Converted: After converting nodes to a cluster (clicking Convert to Cluster) or enabling High Availability mode (clicking Join HA) and before rebooting the node(s).

NOTE:

If you experience problems because the nodes of the HA cluster do not find each other during system startup, navigate to Basic Settings > High Availability and select HA (Fix current). That way the IP address of the HA interfaces of the nodes will be fix, which helps if the HA connection between the nodes is slow.

The DRBD status field indicates whether the latest data (including SPS configuration, audit trails, log files, and so on) is available on both SPS nodes. The primary node (this node) must always be in consistent status to prevent data loss. Inconsistent status means that the data on the node is not up-to-date, and should be synchronized from the node having the latest data.

The DRBD status field also indicates the connection between the disk system of the SPS nodes. The following statuses are possible:

  • Connected: Both nodes are functioning properly.

  • Connected (Disk Failure): A hard disk of the secondary node is not functioning properly and must be replaced. To request a replacement hard disk and for details on replacing the hard disk, contact our Support Team.

  • Invalidated: The data on one of the nodes is considered out-of-sync and should be updated with data from the other node. This state usually occurs during the recovery of a split-brain situation when the DRBD is manually invalidated.

  • Sync source or Sync target: One node (Sync target) is downloading data from the other node (Sync source).

    When synchronizing data, the progress and the remaining time is displayed in the System monitor.

    Caution:

    When the two nodes are synchronizing data, do not reboot or shutdown the primary node. If you absolutely must shutdown the primary node during synchronization, shutdown the secondary node first, and then the primary node.

  • Split brain: The two nodes lost the connection to each other, with the possibility of both nodes being active nodes (that is, primary nodes) for a time.

    Caution:

    Hazard of data loss In this case, valuable audit trails might be available on both SPS nodes, so special care must be taken to avoid data loss. For details on solving this problem, see Recovering from a split brain situation.

  • WFConnection: One node is waiting for the other node, the connection between the nodes has not been established yet.

If a redundant heartbeat interface is configured, its status is also displayed in the Redundant Heartbeat status field, and also in the HA > Redundant field of the System monitor. For a description of redundant heartbeat interfaces, see Redundant heartbeat interfaces.

The possible status messages are explained below.

  • NOT USED: There are no redundant heartbeat interfaces configured.

  • OK: Normal operation, every redundant heartbeat interface is working properly.

  • DEGRADED-WORKING: Two or more redundant heartbeat interfaces are configured, and at least one of them is functioning properly. This status is displayed also when a new redundant heartbeat interface has been configured, but the nodes of the SPS cluster has not been restarted yet.

  • DEGRADED: The connection between the redundant heartbeat interfaces has been lost. Investigate the problem to restore the connection.

  • INVALID: An error occurred with the redundant heartbeat interfaces. Contact the One Identity Support Team for help. For assistance, contact our Support Team.

Recovering One Identity Safeguard for Privileged Sessions (SPS) if both nodes broke down

It can happen that both nodes break down simultaneously (for example because of a power failure), or the secondary node breaks down before the original primary node recovers.

NOTE:

As of One Identity Safeguard for Privileged Sessions (SPS) version 2.0.2, when both nodes of a cluster boot up in parallel, the node with the 1.2.4.1 HA IP address will become the primary node.

To properly recover SPS

  1. Power off both nodes by pressing and releasing the power button.

    Caution:

    Hazard of data loss If SPS does not shut off, press and hold the power button for approximately 4 seconds. This method terminates connections passing SPS and might result in data loss.

  2. Power on the node that was the primary node before SPS broke down. Consult the system logs to find out which node was the primary node before the incident: when a node boots as primary node, or when a takeover occurs, SPS sends a log message identifying the primary node.

    TIP:

    Configure remote logging to send the log messages of SPS to a remote server where the messages are available even if the logs stored on SPS become unaccessible. For details on configuring remote logging, see System logging, SNMP and e-mail alerts.

  3. Wait until this node finishes the boot process.

  4. Power on the other node.

相关文档

The document was helpful.

选择评级

I easily found the information I needed.

选择评级