IBM Support

Removing a server from a Power Enterprise Pool (PEP)

Troubleshooting


Problem

Removing a server from a Power Enterprise Pool (PEP)

Symptom

Power Enterprise Pools allow the administrator flexibility to control workloads and ever changing needs in the environment involving server relocation/removal or addition of new server hardware.

Cause

This document provides the correct procedure for removing servers from a pool when hardware becomes decommissioned, servers are no longer wanted to be sharing in the pool, or server relocation to a new data center.

Environment

HMC Version 7 Release 7.8.0

HMC Version 8 All Releases
HMC Version 9 All Releases

Diagnosing The Problem

If the server is physically removed before removing the server from the pool, later attempts to remove the server or server's resources from the pool will result in the following error(s):

HSCL914F There are no HMCs in the pool that are currently managing server <Server MTMS>

The managed system must be physically reconnected and added back to the HMC in order to complete the procedure outlined in this document. If the servers cannot be added back to the HMC contact IBM Software Support for further assistance.

HSCL915D Mobile CoD resources cannot be removed from server <Server MTMS> because there are no HMCs in the pool that are currently managing the server. Use the force option if you want to remove Mobile CoD resources from the server now.

Resolving The Problem

Procedure for Removing a Server from a Power Enterprise Pool (PEP)

After the Power Enterprise Pool is created, the administrator can update the configuration of the pool at any time by notifying IBM through the PEP office email pcod@us.ibm.com and submitting an addendum to their pool. The submitted request will result in a new configuration file for the pool that will be posted to the CoD website to accomplish configuration changes such as:


1) Adding Mobile Capacity on Demand (CoD) resources to the pool or removing Mobile CoD resources from the pool.
2) Converting permanently activated resources on any server in the pool to Mobile CoD resources.
3) Adding servers to the pool or removing servers from the pool.



Before removing the systems from a pool, all assets (including Mobile CoD resources) that were originally purchased with the system must be returned to that same system serial number. Mobile assets that belong to a system might qualify for transfer to another system serial number depending on normal qualifying guidelines and may require more administrative action.

Systems that are removed from a pool can join another pool and contribute Mobile CoD resources to the new pool and use Mobile CoD resources of other systems. Mobile CoD resources require a pool ID to be recognized

Specifically these following requirements must be met in order to successfully complete this change once the new XML file is posted:

 
  • Ensure that any server to be removed from the pool is in the Standby state or the Operating state. If the state is not one of the listed states, resolve any connectivity or system state issues before proceeding.
    • lssyscfg -r sys -Fname:state
  • Ensure that any server to be removed from the pool does not have any Mobile CoD resource assignments or any unreturned Mobile CoD resources.
    • lscodpool -p pool0487 --level sys

Example result:

name=hmcfsp01,mtms=9179-MHD*1012345,state=Operating,mobile_procs=5,non_mobile_procs=5,unreturned_mobile_procs=0,inactive_procs=6,installed_procs=16,mobile_mem=2048,non_mobile_mem=19456,unreturned_mobile_mem=0,inactive_mem=109568,
installed_mem=131072

In the above example, there are mobile_procs and mobile_mem being used by the server that need to be returned.

Before proceeding, in use pool resources must be returned from usage in the system. The resources can be released through the following methods:

1. Remove the amount of in use pool resources from one or more partitions.
2. Delete partition(s) holding the pool resources
3. Migrate partitions off the system
Resources can be removed from running partitions using dynamic logical partitioning.
Memory and CPU can be removed from "Not activated" partitions using the UI or CLI.  In the UI, set the memory/virtual processors and processing units or dedicated processor values all to 0 and click Save.
image 5889
This is an example of setting cpu to zero on shutdown shared processor partition using command line:
chhwres -m <server> -p <partition name> -o s --procs 0 --procunits 0

Once the resources are released, from the enhanced GUI interface, navigate to the Resources -> All Power Enterprise Pools menu to manage the pool. Select "Processor Resources", then click "Edit".



Change the value of "5" (for example) to "0" to return and click "Save"



The result should show a "0" now for Proposed and Inactive "11", the "6" inactive plus the "5" returned to the pool.



From the left, now navigate to the memory menu, Select "Memory Resources", then click "Edit".



Repeat the steps for the processor, but for memory. Change the value of "2" (for example) to "0" to return and click "Save". The result should show a "0" now for Proposed and Inactive "109", the "107" inactive plus the "2" returned to the pool.



The server can now be removed from the pool by updating the pool information:

1. Apply the new xml configuration file posted on the CoD website: chcodpool -o update -p pool0487 -f pool0487.xml

Proceed with additional actions depending on the end goal for the server.

1. If the server is to be removed from the HMC physically use the rmsysconn command:

Example syntax: rmsysconn -o remove -m mySys

rmsysconn -o remove -m hmcfsp01

Unplug the cables to finish the process of removing the server from the HMC.

2. If the server is only being removed from the pool but remaining on the HMC:

Verify if any activations are needed for permanent, trial, on/off, etc to satisfy workload requirements. To order any new CoD enablement features contact your IBM® Business Partner or IBM sales representative.

3. If the server is being removed from one pool and added to another pool (pool0333 for example) on the same HMC:

Apply the new xml configuration file posted on the CoD website for the other pool: chcodpool -o update -p pool0333 -f pool0333.xml

NOTE: The managed system used for the resource entitlements for the entire pool should be verified by the administrator before sending the addendum to pcod@us.ibm.com. If the managed system being removed is the one for the entitlements, the addendum should included a request to move the entitlements to another server in the pool. Otherwise the following error will be returned:

HSCL90BB Managed system hmcfsp01 cannot be removed from the Power enterprise pool because it has the current Mobile CoD memory activation code for the pool. Contact your CoD administrator.


Check the pool.xml file to determine which server holds the pool entitlements for MemoryResourceEntitlement and ProcessorResourceEntitlement

<MemoryResourceEntitlement><ActivationCode>xxxxxxxxxxxxxxxxxxx</ActivationCode><MTMS>9080-MME*21XXXXX</MTMS></MemoryResourceEntitlement><ProcessorResourceEntitlement><ActivationCode>xxxxxxxxxxxxxxxxxxx</ActivationCode><MTMS>9080-MME*21XXXXX</MTMS></ProcessorResourceEntitlement>
 
Workaround procedure for a removed server from the HMC before removing the server from the pool:
Note:  The pcod team may require proof that the server is no longer in use before granting the updated file.
This procedure must be done from the current pool master.

1) Log in as hscpe
2) Record the pool id and name
lscodpool --level pool
Record the "id" value for use on the runsig command.
Record the "name" value for use on the chcodpool command

3) Delete the pool
Run the following command (HMC V8R8.5 SP1 or later required):
runsig -s 371 -p <pool ID>
where <pool id> is the id value recorded in step 2.
(e.g. runsig -s 371 -p 0195)

4) Confirm pool is deleted.
lscodpool --level pool
If the pool has been deleted (is no longer listed in output) continue to step 5.
If the HMC is an early Version 8 fix level and the pool is still listed, the hmc must be restarted. 
hmcshutdown -r -t now

5) Recover the pool with the new xml file.
a) Login as hscroot (or another hmcsuperadmin user).
b) Copy the new xml file to USB flash or scp the file to the HMC.  
c) Recover the Power enterprise pool using the updated configuration file obtained from IBM pcod
In this example pool1 is the pool name recorded in step 2.  pool0195.xml is the updated xml file obtained from the IBM cod website.  The additional managing HMCs for the pool in the example are hmc2 and hmc3, (you will be prompted to enter the password for each HMC). 

This example shows USB media. If scp was used, provide the path the file was copied to in the -f parameter.

Insert the usb drive.

lsmediadev (to obtain mount points)

mount /media/sdb1

chcodpool -o recover -p pool1 -f /media/sdb1/pool0195.xml -v -a "\"mcs=hmc1/hscroot/,hmc2/hscroot/\""

Enter the hscroot password for the remote hmcs when prompted.  

Review the message.  Verify all existing servers have been added to the pool and correct memory/cpu activation code(s) were successfully entered (HSCL0344/HSCL0345). 

[{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SGGSNP","label":"Hardware Management Console V9"},"ARM Category":[{"code":"a8m0z000000bowEAAQ","label":"Hardware Management Console"}],"ARM Case Number":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Version(s)","Line of Business":{"code":"LOB08","label":"Cognitive Systems"}}]

Document Information

Modified date:
22 September 2021

UID

nas8N1022558

Page Feedback