-
Notifications
You must be signed in to change notification settings - Fork 51
Configure zOSMF HA for Zowe in Sysplex
z/OSMF HA should be configured in Hot Standby mode to ensure availability of REST services. The goal of this configuration is to ensure that one z/OSMF server is always available to provide the REST services. It means that there should be at least one backup (hot-standby) server and a preferred target server. Both targets are active, and both z/OSMF servers are bound to the DVIPA. So, preferred z/OSMF server receives all new incoming requests. If it fails or the system becomes down, new requests are routed to the backup (hot-standby) server. In this mode, the distributing DVIPA does not perform load balancing of requests across multiple systems.
- The SAF database is to be shared between multiple systems, the device on which the SAF database resides must be shared.
TSO RVARY
command will show you if SAF database is in DATA SHARING MODE. - Shared File System environment https://www.ibm.com/support/knowledgecenter/en/SSLTBW_2.4.0/com.ibm.zos.v2r4.bpxb200/uxsys.htm
- JESPLEX/JES2 Multi-Access Spool (MAS) environment
- TCP/IP availability by Sysplex Distributor (DVIPA)
- Extended MCS console (EMCS)
- If it is a first-time z/OSMF setup, please follow the link below to create z/OSMF nucleus on your system: https://www.ibm.com/support/knowledgecenter/SSLTBW_2.4.0/com.ibm.zos.v2r4.izua300/izulite_CreateTheNucleus.htm
- z/OSMF security authorizations will be used by all z/OSMF servers across multiple systems.
- The different STARTED profiles should be defined for z/OSMF servers.
- Create the different shared file system for each z/OSMF server.
- Each system uses a different IZUPRMxx member. For example IZUPRM0A and IZUPRM0B
-
AUTOSTART_GROUP, more than one z/OSMF server (independent z/OSMF instances) is to be autostarted in a Sysplex. For instance, System A will autostart a server and similarly, System B will autostart the second z/OSMF server.
- z/OSMF has a default autostart group (IZUDFLT) which is used in monoplex or single z/OS image. To have more z/OSMF servers autostarted in a Sysplex, you must associate each server and the systems it serves with a unique autostart group name. For example, AUTOSTART_GROUP('IZUDFLA') for System A and AUTOSTART_GROUP('IZUDFLB') for System B
- AUTOSTART(LOCAL) should be used by all z/OSMF instances.
- HOSTNAME, the DVIPA address will be used as the z/OSMF host name for all instances.
- HTTP_SSL_PORT, all servers are listening on the same port.
- KEYRING_NAME, all servers should use the same key ring such as IZUKeyring.IZUDFLT.
- SERVER_PROC, each z/OSMF server should use the different started procedure. For example IZUSVRA and IZUSVRB.
- ANGEL_PROC, all z/OSMF servers can connect to the same z/OSMF angel process such as IZUANG1.
- SAF_PREFIX, they should use the same SAF profile prefix such as IZUDFLT.
- USER_DIR, each instance uses a shared file system with a unique mount point for each AUTOSTART group that be automatically started. For example, /global/zosmf/zosmfa and /global/zosmf/zosmfb
The following DVIPA configuration ensures the availability of z/OSMF for Hot-Standby. For example, suppose that you have a Sysplex of two systems: A, B:
- Enable dynamic XCF on each host:
-
IPCONFIG SYSPLEXROUTING DYNAMICXCF x.x.x.A 255.255.255.0 1
for SYSA -
IPCONFIG SYSPLEXROUTING DYNAMICXCF x.x.x.B 255.255.255.0 1
for SYSB
- Define a dynamic VIPA (DVIPA) for both systems:
VIPADYNAMIC
VIPADEFINE 255.255.255.0 x.x.x.V
VIPADISTRIBUTE DEFINE DISTM HOTSTANDBY x.x.x.V
PORT 443 DESTIP
x.x.x.A PREFERRED
x.x.x.B BACKUP
ENDVIPADYNAMIC
Note:
- x.x.x.A , the home address for SYSA
- x.x.x.B , the home address for SYSB
- x.x.x.V , Dynamic VIP Address
The VIPADISTRIBUTE statement with PREFERRED and BACKUP settings is used to enable automatic dynamic VIPA takeover to occur, if needed. Both z/OSMF servers are bound to the DVIPA x.x.x.V. With both z/OS systems active in the Sysplex, the preferred z/OSMF server, SYSA receives all new incoming requests. If the SYSA fails, new work requests for z/OSMF are routed to the server on SYSB. When SYSA resumes normal operations, new work requests for z/OSMF are routed to SYSA again. This behavior occurs because AUTOSWITCHBACK option of VIPADISTRIBUTE statement is in effect by default.
If AUTOSWITCHBACK option is in place, the new TSO REST API or Console REST API requests might fail if you have already a TSO address space or an Extended MCS console established/created on SYSB (in other TSO/E address space). This issue happens because REST Console API and REST TSO API are stateful, and intend to reuse undercover TSO address space and EMCS console for same user. Coding NOAUTOSWITCHBACK option in VIPADISTRIBUTE statement might resolve this issue. NOAUTOSWITCHBACK specifies that you do not want the distributor to switch back to the preferred target when it becomes available.
VIPADYNAMIC
VIPADEFINE 255.255.255.0 x.x.x.V
VIPADISTRIBUTE DEFINE DISTM HOTSTANDBY NOAUTOSWITCHBACK x.x.x.V
PORT 443 DESTIP
x.x.x.A PREFERRED
x.x.x.B BACKUP
ENDVIPADYNAMIC
- Configure zOSMF HA for Zowe in Sysplex
- Enable JWT function on zOSMF HA servers
- Enable single sign-on on zOSMF HA servers
- Enable Zowe to generate and evaluate PassTickets for APIML Services Zowe HA
- Deploy Zowe in Sysplex
- Test Zowe in Sysplex
- List of changes to the current documentation
- Additions to the current documentation