Rolling Reload of NonStop Kernel Processors

ManualsBrandsHP ManualsServerHP Integrity NonStop Service Information

Rolling Reload of NonStop Kernel Processors 2/18/04

Hewlett-Packard Development Company L.P. 1 of 7 CSSI Website

A Procedure for Rolling Reload of NonStop Kernel

Processors

Rolling reloads can be used to avoid system loads in some situations such as processor

upgrades and certain software product revision (SPR) installations. However, care must

be exercised when performing a rolling reload to avoid causing an outage. To minimize

the chance of problems during or after a rolling reload, follow the guidelines and

procedures presented in this document.

1 Rolling reload considerations

1.1 Rolling reload vs. system load

Many HP NonStop system customers successfully use rolling reloads and other on-line

software installation techniques. However, problems can potentially occur when halting

and reloading processors because of unexpected system or application behavior, load

balancing, and non-standard configurations. If the next processor is halted before the

previously halted processor has fully recovered, an outage is probable. Rolling reloads

should be used only when all of the following conditions are true:

• It is not possible to schedule an outage for an SPR installation.

• The SPR(s) being installed is very important - for example, an SPR recommended

by an Outage Prevention Notice (OPN).

• The SPR documentation states that the SPR can be installed with a rolling reload.

• The application and system configuration are well understood and fault-tolerant.

• It is possible to load balance the application after repeated processor halts and

reloads.

Note: The time needed to perform a rolling reload is usually longer than the time needed

to perform a system load. If the application must be halted during the rolling reload, you

can perform a system load instead for a potentially shorter application outage.

1.2 Application considerations

Successful application recovery from a rolling reload requires fault-tolerant application(s)

that can be quickly rebalanced following a processor halt and reload.

If your application does not meet these criteria and it is possible to have an application

interruption (but not a system load), an option is shutting down the application,

performing the rolling reload, and then restarting the application. Note, however, that the

resulting application interruption will probably be longer than that from a system load.

1.3 A few other tips

• Whenever possible, test the rolling reload on a development system using the

production application and expected production workload.

• Perform the rolling reload when the system has its minimal utilization.

• Carefully follow the steps described in the remainder of this document.

Summary of content (7 pages)

PAGE 1
Rolling Reload of NonStop Kernel Processors 2/18/04 A Procedure for Rolling Reload of NonStop Kernel Processors Rolling reloads can be used to avoid system loads in some situations such as processor upgrades and certain software product revision (SPR) installations. However, care must be exercised when performing a rolling reload to avoid causing an outage. To minimize the chance of problems during or after a rolling reload, follow the guidelines and procedures presented in this document.
PAGE 2
Rolling Reload of NonStop Kernel Processors 2/18/04 2 Planning a rolling reload. Rolling reloads are significant system events and require careful planning. The first step is to gather and evaluate all documentation pertaining to the planned update, such as SPR documentation or Hotstuff messages. The Global Customer Support Center (GCSC) is an excellent resource for rolling reload assistance. Contact the GCSC after completing a rolling reload plan.
PAGE 3
Rolling Reload of NonStop Kernel Processors 2/18/04 (3) reload each processor with a full recovery by each process, and (4) perform system and application recovery after all processors are reloaded. The following text describes how to determine the amount of time needed for each of these steps. (1) Determine how much time is required for backup processes to fully takeover.
PAGE 4
Rolling Reload of NonStop Kernel Processors 2/18/04 3 Rolling reload preparation This section describes the activities that should be performed prior to starting the rolling reload. 3.1 Prepare the SPRs to be installed Code to be updated should be present on the preferred load subvolume (often $SYSTEM.SYSnn, $SYSTEM.SYSTEM, or other $SYSTEM.Z*) depending on subsystem installation methods. For the best results, use DSM/SCM Build/Apply.
PAGE 5
Rolling Reload of NonStop Kernel Processors 2/18/04 4 Rolling reload implementation Before starting the rolling reload, complete the procedures described in Sections 2 and 3. Verify the fault-tolerance of the system, paying particular attention to the ServerNet Fabrics and processor-to-ServerNet addressable controller (SAC) relationships. To implement a rolling reload, perform the pre-checks in Section 4.1, halt and reload each processor one at a time using the procedures in Sections 4.2-4.
PAGE 6
Rolling Reload of NonStop Kernel Processors • • • 2/18/04 Perform similar checks for applications and other subsystems such as ServerNet/FX and Telco-specific SACs. If necessary, verify the health of the applications. Verify that the time since the halt is about what was determined in the planning phase (Section 2.5, Determine how much time is required for a rolling reload). 4.3 Execute update step(s), if needed before reload.
PAGE 7
Rolling Reload of NonStop Kernel Processors • • • • • • 2/18/04 Verify that all SLSA SACs are in the STARTED state, not STARTING. Check the access from both fabrics and all processors configured for each SAC. Verify that there are no new OSM/TSM alarms. Perform similar checks for applications and other subsystems such as ServerNet/FX and Telco-specific SACs. If necessary, verify the health of the applications. If necessary, verify that the updates have been installed and are active on this processor.