IBM TS7700 implementation
This chapter describes how to implement the IBM TS7700 on IBM Z platforms. From a software perspective, differences exist between the TS7700C, TS7760D, TS7760T, TS7740, TS7720D, and the TS7720T. If no specific differences are indicated, the implementation steps apply to all models. Otherwise, the differences are explained in each relevant step.
This chapter includes the following topics:
For more information about defining a tape subsystem in a DFSMS environment, see IBM TS3500 Tape Library with System z Attachment A Practical Guide to Enterprise Tape Drives and TS3500 Tape Automation, SG24-6789, and IBM TS4500 R5 Tape Library Guide, SG24-8235.
6.1 TS7700 implementation
The following sections describe the implementation and installation tasks to set up theTS7700. Specific names are used in this chapter if a certain task applies only to one of the six models because there are slight differences between theTS7760C, TS7760D, TS7760T, TS7740, TS7720D, and TS7720T.
Because the TS7720D, TS7760C, and TS7760D do not have a tape library that is attached, the implementation steps that are related to a physical tape library for IBM TS4500 or IBM TS3500 do not apply.
You can install the TS7760T, TS7740, or TS7720T together with your existing TS3500 tape library, or install them with a new TS4500 to serve as the physical back-end tape library. When using a TS3500 as the physical back-end tape library to either the TS7760T, TS7740, or TS7720T, the IBM 3953 Library Manager is no longer required because the Library Manager functions are provided by the TS7700 Licensed Internal Code.
6.1.1 Implementation tasks
The TS7700 implementation can be logically separated into three major sections:
•TS7760T, TS7740, or TS7720T and tape library setup
Use the TS7760T, TS7740, or TS7720T and the TS4500/TS3500 tape library interfaces for these setup steps:
– Defining the logical library definitions of the TS7760T, TS7740, or TS7720T, such as physical tape drives and cartridges by using the TS4500/TS3500 tape library GUI, which is the web browser interface with the TS4500/TS3500 tape library.
– Defining specific settings, such as encryption, and inserting logical volumes into the TS7760T, TS7740, or TS7720T. You can also use the Management Interface (MI) to define logical volumes, management policies, and volume categories.
This chapter provides details about these implementation steps.
•TS7700 hardware input/output (I/O) configuration definition
This section relates to the system generation. It consists of processes, such as FICON channel attachment to the host, hardware configuration definition (HCD) or input/output configuration program (IOCP) definitions, and missing-interrupt handler (MIH) settings. This activity can be done before the physical hardware installation and it can be part of the preinstallation planning.
•TS7700 software definition
This is where you define the new virtual tape library to the individual host operating system. In an IBM Z environment with Data Facility Storage Management Subsystem (DFSMS)/IBM MVS, this includes updating DFSMS automatic class selection (ACS) routines, object access method (OAM), and your tape management system (TMS) during this phase. You also define Data Class (DC), Management Class (MC), Storage Class (SC), and Storage Group (SG) constructs and selection policies, which are passed to the TS7700.
These three groups of implementation tasks can be done in parallel or sequentially. HCD and host definitions can be completed before or after the actual hardware installation.
6.2 TS4500/TS3500 tape library definitions
Use this section if you are implementing theTS7760T, TS7740, or TS7720T with a TS4500/TS3500 tape library in an IBM Z environment. If your TS7700 does not have an associated tape library (TS7760C, TS7760D, or TS7720D), see 6.3, “Setting up the TS7700” on page 230.
Your IBM Service Support Representative (IBM SSR) installs the TS7760T, TS7740, or TS7720T hardware, its associated tape library, and the frames. This installation does not require your involvement other than the appropriate planning. For more information, see Chapter 4, “Preinstallation planning and sizing” on page 135.
|
Clarification: The steps that are described in this section relate to the installation of a new IBM TS4500/TS3500 tape library with all of the required features, such as Advanced Library Management System (ALMS), installed. If you are attaching an existing IBM TS3500 tape library that is attached to Open Systems hosts to IBM Z hosts, see IBM TS3500 Tape Library with System z Attachment A Practical Guide to Enterprise Tape Drives and TS3500 Tape Automation, SG24-6789, or IBM TS4500 R5 Tape Library Guide, SG24-8235 for more information about extra steps that might be required.
|
The following tasks are for TS4500/TS3500 library definition. For the detailed procedure, see 9.5.1, “The tape library with the TS7700T cluster” on page 524.
•Defining a logical library
– Ensuring that ALMS is enabled
– Creating a logical library with ALMS
– Setting the maximum cartridges for the logical library
•Adding drives to the logical library
Each tape-attached TS7700 will be associated to only one logical library in the physical library, and each logical library can be associated to only one TS7700. TS7700T will require a minimum of four tape drives installed to be operational, and a maximum of sixteen can be used.
•Defining control path drives
Each TS7760T, TS7740, or TS7720T requires the definition of four tape drives (among installed ones) as control path drives.
•Defining the Encryption Method for the new logical library
TS7700T only supports encryption when the logical library has been configured to use the System-Managed encryption method
•Defining CAPs (Cartridge Assignment Policies)
•Inserting TS7760T, TS7740, or TS7720T physical volumes
•Assigning cartridges in the TS4500/TS3500 tape library to the logical library partition
This procedure is necessary only if a cartridge was inserted, but a Cartridge Assignment Policy (CAP) was not provided in advance.
6.3 Setting up the TS7700
This section describes the tasks that are required to set up the TS7700.
6.3.1 Definitions for TS7760T, TS7740, or TS7720T
If you have a TS7760C, TS7760D, or TS7720D, skip to 6.3.2, “TS7700 definitions” on page 230. For the more information about procedures, see 9.5.2, “TS7700T definitions” on page 542.
The tasks that are listed in this section are for TS7760T, TS7740, or TS7720T only:
•Defining VOLSER ranges for physical volumes
•Defining physical volume pools:
– Reclaim threshold setting
– Inhibit Reclaim schedule
6.3.2 TS7700 definitions
Use the TS7700 MI to continue with the following TS7700 Virtualization subsystem setup tasks. For more information, see 9.5.3, “TS7700 definitions” on page 552:
•Defining scratch categories and logical volume expiration time
•Defining TS7700 constructs
To use the Outboard Policy Management functions, you must define four constructs:
– Storage Group (SG)
– Management Class (MC)
– Storage Class (SC)
– Data Class (DC)
•TS7700 licensing
•Defining ISKLM (IBM Security Key Lifecycle Manager) addresses for external management of encryption keys
•Defining Simple Network Management Protocol (SNMP)
•Defining rocket-fast system for log processing (RSYSLOG)
•Cluster Network Settings
•Enabling Internet Protocol Security (IPSec)
•Security Settings
•Inserting logical virtual volumes
•Cloud Tier Settings (TS7760C only)
To useTS7760C Cloud Storage Tier, you must define three settings:
– Cloud pools
– Cloud accounts
– Containers, which also requires:
• Cloud URL
• Cloud URL to cluster association
|
Note: For more information about tasks that required for implementation of a TS7760C, see IBM TS7760 R4.2 Cloud Storage Tier Guide, REDP-5514:
|
6.4 Hardware configuration definition
This section describes the process of defining the TS7700 through the HCD interface. Usually, HCD definitions are made by IBM z/OS system administrators. A helpful approach is to complete a table with all of the definitions that the administrators will need, and then give the table to the administrators.
Table 6-1 is an example definition table for a stand-alone cluster. In general, all of the blank cells must be completed by system administrators because they know what channels are free, what control unit (CU) numbers are free, and so on.
Table 6-1 HCD definitions table for Cluster 0
|
CHPID
|
CU
|
CUADD
|
Link
|
Devices
|
ADD
|
LIB-ID
|
Libport
|
|
|
|
0
|
|
|
00-0F
|
|
01
|
|
|
|
1
|
|
|
00-0F
|
|
02
|
|
|
|
2
|
|
|
00-0F
|
|
03
|
|
|
|
3
|
|
|
00-0F
|
|
04
|
|
|
|
4
|
|
|
00-0F
|
|
05
|
|
|
|
5
|
|
|
00-0F
|
|
06
|
|
|
|
6
|
|
|
00-0F
|
|
07
|
|
|
|
7
|
|
|
00-0F
|
|
08
|
|
|
|
8
|
|
|
00-0F
|
|
09
|
|
|
|
9
|
|
|
00-0F
|
|
0A
|
|
|
|
A
|
|
|
00-0F
|
|
0B
|
|
|
|
B
|
|
|
00-0F
|
|
0C
|
|
|
|
C
|
|
|
00-0F
|
|
0D
|
|
|
|
D
|
|
|
00-0F
|
|
0E
|
|
|
|
E
|
|
|
00-0F
|
|
0F
|
|
|
|
F
|
|
|
00-0F
|
|
10
|
|
|
|
10
|
|
|
00-0F
|
|
11
|
|
|
|
11
|
|
|
00-0F
|
|
12
|
|
|
|
12
|
|
|
00-0F
|
|
13
|
|
|
|
13
|
|
|
00-0F
|
|
14
|
|
|
|
14
|
|
|
00-0F
|
|
15
|
|
|
|
15
|
|
|
00-0F
|
|
16
|
|
|
|
16
|
|
|
00-0F
|
|
17
|
|
|
|
17
|
|
|
00-0F
|
|
18
|
|
|
|
18
|
|
|
00-0F
|
|
19
|
|
|
|
19
|
|
|
00-0F
|
|
1A
|
|
|
|
1A
|
|
|
00-0F
|
|
1B
|
|
|
|
1B
|
|
|
00-0F
|
|
1C
|
|
|
|
1C
|
|
|
00-0F
|
|
1D
|
|
|
|
1D
|
|
|
00-0F
|
|
1E
|
|
|
|
1E
|
|
|
00-0F
|
|
1F
|
6.4.1 Defining devices through HCD
You can define up to 31 CUs with 16 devices each per cluster in the grid configuration. Use CUADD=0 - CUADD=7 and LIBPORT-IDs of 01 - 08 for the first eight CUs, as shown in Table 6-2.
Table 6-2 CUADD and LIBPORT-ID for the first set of 256 virtual devices
|
CU
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
8
|
|
CUADD
|
0
|
1
|
2
|
3
|
4
|
5
|
6
|
7
|
|
LIBPORT-ID
|
01
|
02
|
03
|
04
|
05
|
06
|
07
|
08
|
For the ninth to sixteenth CUs, use CUADD=8 - CUADD=F and LIBPORT-IDs of 09 - 10, as shown in Table 6-3.
Table 6-3 CUADD and LIBPORT-ID for the second set of virtual devices
|
CU
|
9
|
10
|
11
|
12
|
13
|
14
|
15
|
16
|
|
CUADD
|
8
|
9
|
A
|
B
|
C
|
D
|
E
|
F
|
|
LIBPORT-ID
|
09
|
0A
|
0B
|
0C
|
0D
|
0E
|
0F
|
10
|
Figure 6-1 and Figure 6-2 on page 234 show the two important windows for specifying a tape CU. To define devices by using HCD, complete the following steps:
|
------------------------ Add Control Unit ---------------------
CBDPCU10 Specify or revise the following values. Control unit number . . . . 0440 + Control unit type . . . . . 3490 + Serial number . . . . . . . __________ Description . . . . . . . . ________________________________ Connected to switches . . . 01 01 01 01 __ __ __ __ + Ports . . . . . . . . . . . D6 D7 D8 D9 __ __ __ __ + If connected to a switch: Define more than eight ports . 2 1. Yes 2. No Propose CHPID/link addresses and unit addresses. . . . . . . . .2 1. Yes
2. No F1=Help F2=Split F3=Exit F4=Prompt F5=Reset F9=Swap F12=Cancel |
Figure 6-1 Add the first TS7700 CU through HCD (Part 1 of 2)
2. The window that is shown in Figure 6-2 opens. Select the processor to which the CU is to be connected.
|
-------------------------- Add Control Unit -------------------------
CBDPCU12 Specify or revise the following values. Control unit number . : 0440 Type . . . . . . : 3490 Processor ID . . . . . : PROC1 This is the main processor Channel Subsystem ID . : 0 Channel path IDs . . . . 40 50 60 70 __ __ __ __ + Link address . . . . . . D6 D7 D8 D9 __ __ __ __ + Unit address . . . . . . 00 __ __ __ __ __ __ __ + Number of units . . . . 16 ___ ___ ___ ___ ___ ___ ___ Logical address . . . . 0 + (same as CUADD) Protocol . . . . . . . . __ + (D,S or S4) I/O concurrency level . 2 + (1, 2 or 3) F1=Help F2=Split F4=Prompt F5=Reset F9=Swap F12=Cancel |
Figure 6-2 Add the first TS7700 CU through HCD (Part 2 of 2)
|
Tip: When the TS7700 is not attached through Fibre Channel connection (FICON) directors, the link address fields are blank.
|
3. Repeating the previous process, define the second through the 16th TS7700 virtual tape CUs, specifying the logical unit address (CUADD)=1 - F, in the Add Control Unit windows. The Add Control Unit summary window is shown in Figure 6-2.
4. To define the TS7700 virtual drives, use the Add Device window that is shown in Figure 6-3.
|
------------------------------- Add Device ---------------------------
CBDPDV10 Specify or revise the following values. Device number . . . . . . . . 0A40 (0000 - FFFF) Number of devices . . . . . . 16__ Device type . . . . . . . . . 3490_________ + Serial number . . . . . . . . __________ Description . . . . . . . . . ________________________________ Connected to CUs . . 0440 ____ ____ ____ ____ ____ ____ ____ + F1=Help F2=Split F3=Exit F4=Prompt F5=Reset F9=Swap F12=Cancel |
Figure 6-3 Add the first 16 drives through HCD
5. After you enter the required information, you can specify to which processors and operating systems the devices are connected to. Figure 6-4 shows the window that is used to update the processor’s view of the device.
|
------------------------ Define Device / Processor---------------------------
CBDPDV12 Specify or revise the following values. Device number . : 0A40 Number of devices . . . . : 16 Device type . . : 3490 Processor ID. . : PROC1 This is the main processor
Unit address . . . . . . . . . . 00 +(only necessary when different from the last 2 digits of device number) Time-Out . . . . . . . . . . . . No (Yes or No)
STADET . . . . . . . . . . . . . No (Yes or No) Preferred CHPID . . . . . . . . __ + Explicit device candidate list . No (Yes or No) F1=Help F2=Split F4=Prompt F5=Reset F9=Swap F12=Cancel |
Figure 6-4 HCD Define Device / Processor window
6. After you enter the required information and specify to which operating systems the devices are connected, the window in Figure 6-5 is displayed, where you can update the device parameters.
|
CBDPDV13 Define Device Parameters / Features Row 1 of 6
Command ===> __________________________________________ Scroll ===> PAGE
Specify or revise the values below.
Configuration ID . : AB MVS operating system
Device number . . : 0440 Number of devices :16
Device type . . . : 3490
Parameter /
Feature Value P Req. Description
OFFLINE Yes Device considered online or offline at IPL
DYNAMIC Yes Device supports dynamic configuration
LOCANY No UCB can reside in 31 bit storage
LIBRARY Yes Device supports auto tape library
AUTOSWITCH No Device is automatically switchable
LIBRARY-ID CA010 5-digit library serial number
LIBPORT-ID 01 2 digit library string ID (port number)
MTL No Device supports manual tape library
SHARABLE No Device is Sharable between systems
COMPACT Yes Compaction
***************************** Bottom of data ****************************
F1=Help F2=Split F4=Prompt F5=Reset F7=Backward
F8=Forward F9=Swap F12=Cancel F22=Command
|
Figure 6-5 Define Device Parameters HCD window
|
Tips: Consider the following points:
•If you are defining drives that are installed in a system-managed IBM tape library, such as the TS7700, you must specify LIBRARY=YES.
•If more than one IBM Z host will be sharing the virtual drives in the TS7700, specify SHARABLE=YES. This forces OFFLINE to YES. It is up to the installation to ensure the correct serialization from all attached hosts.
•You must use the composite library ID of the TS7700 in your HCD definitions.
•The distributed library IDs are not defined in HCD.
|
To define the remaining TS7700 3490E virtual drives, repeat this process for each CU in your implementation plan.
6.4.2 Activating the I/O configuration
There are differences in the concurrent input/output definition file (IODF) activation process between a new tape library implementation and a configuration change that is made to an existing library. Changes to the virtual devices’ address range of an existing library is an example of where concurrent IODF activation is useful.
As an alternative to the procedures described next, you can always perform an initial program load (IPL) or restart of the system.
Installing a new tape library
If you are installing a TS7700 for the first time, from a host software definition point of view, this is an installation of a new library. When you are activating the IODF for a new tape library, the following steps must be completed to get the tape library or TS7700 online without an IPL of your systems:
1. Activate the IODF.
2. Run MVS console command VARY ONLINE to vary online the devices in the library. This command creates some of the control blocks. You should see the following message:
IEA437I TAPE LIBRARY DEVICE(ddd), ACTIVATE IODF=xx, IS REQUIRED
3. Do the final ACTIVATE. This action is required to build the eligible device table (EDT) for MVS Allocation.
After activation, you can check the details by using the DEVSERV QTAPE command. See 10.1.2, “MVS system commands” on page 589.
Modifications to an existing tape library
When you are modifying an existing tape library so that existing device addresses can be changed, complete the following steps:
1. Activate an IODF that deletes all devices from the library.
2. Activate an IODF that defines all of the devices of the modified library.
3. Run MVS console command VARY ONLINE to vary online the devices in the library. This creates some of the control blocks. You see the following message:
IEA437I TAPE LIBRARY DEVICE(ddd), ACTIVATE IODF=xx, IS REQUIRED
4. Do the final ACTIVATE.
Alternatively, you can use the DS QL,nnnnn,DELETE (where nnnnn is the LIBID) command to delete the library’s dynamic control blocks. If you have IODF with LIBID and LIBPORT coded already, and you deleted the library’s dynamic control blocks, complete the following steps:
1. Use QLIB LIST to see whether the INACTIVE control blocks are deleted.
2. Use ACTIVATE IODF to redefine the devices.
3. Use QLIB LIST to verify that the ACTIVE control blocks are properly defined.
If LIBRARY-ID (LIBID) and LIBPORT-ID are not coded, after you delete the library’s dynamic control blocks, complete the following steps:
1. Run MVS console command VARY ONLINE to vary on the devices in the library. This creates some control blocks, and you see the following message:
IEA437I TAPE LIBRARY DEVICE(ddd), ACTIVATE IODF=xx, IS REQUIRED
2. Activate an IODF that defines all of the devices in the modified library.
3. Use QLIB LIST to verify that the ACTIVE control blocks are properly defined.
6.5 Setting values for the Missing Interrupt Handler
The TS7700 emulates 3490E devices and, for clusters running R4.1.2 code levels and later, will communicate a default Primary MIH Time OUT value of 45 minutes to the host operating system in the Read Configuration Data channel control word (CCW X’FA’).
|
Important: An MIH value of 45 minutes is preferable for the virtual devices in a multi-cluster grid when a copy consistency for the remote clusters is set to RUN.
|
The MIH timeout value applies only to the virtual 3490E drives and not to the real IBM TS1150/TS1140/TS1130/TS1120/3592 drives that the TS7740 manages in the back end. The host knows only about logical 3490E devices.
Table 6-4 lists traditionally used MIH values, which can be adjusted depending on specific operational factors.
Table 6-4 Tape device MIH values
|
Tape device
|
MIH
|
|
TS7700 stand-alone grid with 3490E emulation drives
|
20 minutes
|
|
TS7700 multi-cluster grid with 3490E emulation drives
|
45 minutes
|
|
TS7700 multi-cluster grid with 3490E emulation drives and not using Rewind Unload (RUN) copy policies
|
20 minutes
|
If adjustment is needed, the MIH value can be specified in the PARMLIB member IECIOSxx. Alternatively, you can also set the MIH values using the IBM Z operator command SETIOS. A user-specified MIH value will override the default mentioned above, and will be effective until it is manually changed or until the system is initialized.
Use the following statements in PARMLIB, or manual commands to display and set your MIH values:
•You can specify the MIH value in the IECIOSxx PARMLIB member:
MIH DEV=(0A40-0A7F),TIME=45:00
•To manually specify MIH values for emulated 3490E tape drives, use this command:
SETIOS MIH,DEV=(0A40-0A7F),TIME=45:00
– To display the new settings, use this command:
D IOS,MIH,DEV=0A40
– To check the current MIH time, use this command:
D IOS,MIH,TIME=TAPE
For more information about MIH settings, see MVS Initialization and Tuning Reference, SA23-1380.
During IPL (if the device is defined to be ONLINE) or during the VARY ONLINE in process, some devices (such as the IBM 3590/3592 physical tape devices) might present their own MIH timeout values through the primary/secondary MIH timing enhancement that is contained in the self-describing data for the device. The Primary MIH Time OUT value is used for most I/O commands, but the Secondary MIH Time OUT value can be used for special operations, such as long-busy conditions or long-running I/O operations.
Any time that a user specifically sets a device or device class to an MIH timeout value that is different from the default for the device class that is set by IBM, that value overrides the device-established Primary MIH Time OUT value. This implies that if an MIH timeout value that is equal to the MIH default for the device class is explicitly requested, IOS does not override the device-established Primary MIH Time OUT value. To override the device-established Primary MIH Time OUT value, you must explicitly set a timeout value that is not equal to the MIH default for the device class.
6.6 Defining fence actions
In most cases, the TS7700 has excellent capabilities in detecting when it is suffering from problems that could impact its availability and require an external intervention to repair and recover. However, certain type of failures are difficult to detect for the cluster presenting the problem. This difficulty can be caused by the intermittent nature of the failure, the frequency at which it happens, its duration, or other circumstances that can escape the normal mechanism for local detection. Under those circumstances, the failing cluster could still report itself as being in a good-working condition, even if it is not.
This type of failure is known as Sick But Not Dead (SBND). Even if only a single cluster is in this condition, the failure could produce effects that might affect the performance, the functionality, or both of other clusters that are present in the same grid.
Grid-wide consequences for this type of problem can be prevented by a “fence” mechanism that is available when all machines in the grid are running code level R4.1.2 or later. The mechanism isolates the failing box so that it receives the corresponding repair action, while other machines on the grid will keep running without being affected by the unscheduled peer-outage.
In this scenario, the failing box is unable to recognize that it has a problem and keeps trying to take part on grid transactions. Because of this, the other clusters in the grid have the responsibility of recognizing this condition and for executing “remote fence” actions against the sick cluster. These actions can be optionally configured (depending on business requirements) using the LIBRARY REQUEST interface as follows:
•Remote Cluster Fence Function enablement:
LI REQ, <composite-library>, FENCE, <ENABLE|DISABLE>
The default is disable. This control allow/disallow execution of remote cluster fence actions from the perspective of the entire grid.
•Primary fence action:
LI REQ, <distributed-library>, FENCE, ACTION, PRI, <NONE|ALERT|OFFLINE|REBOOT|REBOFF>
This command has these options:
– NONE: This is the default setting. Other clusters in the grid will not be allowed to do remote fencing against the specified cluster (“distributed-library”) if it fails. Therefore, the system depends on the ability of the local cluster to detect that something is wrong with itself.
– ALERT: Other clusters will only surface an alert when they believe that the specified grid member is having problems.
– OFFLINE: Other clusters will trigger a force-offline sequence against the specified cluster to exclude it from grid transactions if the corresponding thresholds are met. If this primary action fails for any reason, a secondary fence action (if enabled) will be initiated.
– REBOOT: Other clusters will trigger a reboot sequence against the specified cluster if the corresponding thresholds are met. In some cases, this process can actually help it to recover when the machine attempts to go online again. If the machine fails to be rebooted for any reason, a secondary fence action (if enabled) will be initiated.
– REBOFF (reboot and stay offline): Other clusters will trigger a reboot sequence against the specified cluster when the corresponding thresholds are met. This sequence will prevent it from attempting to come online again, so it can be subject to deep troubleshooting routines. If the machine fails to be rebooted for any reason, a secondary fence action (if enabled) will be initiated.
•Secondary fence action:
LI REQ, <distributed-library>, FENCE, ACTION, SEC, <ENABLE|DISABLE>
Referred as “Isolate from grid network”, this secondary action will be triggered against the specified cluster only if it was previously enabled (default is “disable”), and selected primary action (only for offline or reboot cases) was detected to fail. In this case, other clusters will automatically execute a reconfiguration of the intercluster network to logically exclude the failing cluster from any grid transactions.
•Enable/Disable Capturing IBM AIX® dump at reboot action:
LI REQ, <distributed-library>, FENCE, ACTION, AIXDUMP, <ENABLE|DISABLE>
An AIX dump is a data capture mechanism that fetches information related to the IBM AIX kernel state at the time it is taken. This information can then be used by authorized IBM Service/Support personnel during the diagnostic process in case of failure. This dump is taken only when enabled (default is disable) and the selected primary fence action is reboot or reboot and stay offline.
•Remote Cluster Fence Thresholds:
LI REQ, <composite-library>, FENCE, THRESHLD, <TMO|ERR|SCRVOAVG|PRIVOAVG|VCAVG|TOKAVG|EVALWIN> <value>
Thresholds describe tolerance levels for detection of SBND conditions. These settings will have grid-wide effect, and there are several options that can be adjusted depending on your business needs:
– Threshold concepts (third keyword in the LI REQ command):
• TMO (timeout event counts): Count cluster handshakes that have reached their configured timeout before a response occurs. This amount relates to the combined sum of timeout incidents for scratch mounts, private mounts, volume close, and token handshakes.
• ERR (error event counts): Count cluster handshakes that have returned an explicit failure. This amount relates to the combined sum of explicit error incidents for scratch mounts, private mounts, volume close, and token handshakes.
• SCRVOAVG (scratch volume open average): The average peer handshake time to open or mount a scratch virtual tape.
• PRIVOAVG (private volume open average): The average peer handshake time to open or mount a private virtual tape.
• VCAVG (volume close average): The average peer handshake time to close or demount a tape, which excludes the RUN (Rewind and UNload) copy duration.
• TOKAVG (token handshake average): The average peer handshake time to process a miscellaneous token handshake.
• EVALWIN (evaluation window): The referred lapse in minutes used to calculate counts and averages for target concepts.
– Threshold values (fourth keyword in the LI REQ command):
• TMO: This setting allows values from 0 to 1000 (default is 20).
• ERR: This setting allows values from 0 to 1000 (default is 20).
• SCRVOAVG: This setting allows values from 0 to 1200 seconds (default is 180 seconds).
• PRIVOAVG: This setting allows values from 0 to 1200 seconds (default is 180 seconds).
• VCAVG: This setting allows values from 0 to 1200 seconds (default is 180 seconds).
• TOKCAVG: This setting allows values from 0 to 180 seconds (default is 120 seconds).
• EVALWIN: This setting allows values from 1 to 30 minutes (default is 7 minutes).
•Showing current Fence Status and Settings:
LI REQ, <composite-library>, FENCE, SHOW
|
Note: For more information about the Host Console Request functions and their responses, see IBM Virtualization Engine TS7700 Series z/OS Host Command Line Request User’s Guide, WP101091:
|
6.7 TS7700 software definitions
This section describes the software definition considerations for implementing the TS7700 in z/OS. The TS7700 must be defined as a new tape library with emulated 3490E Tape Drives from the host system.
To use the TS7700, at least one SG must be created to enable the TS7700 tape library virtual drives to be allocated by the storage management subsystem (SMS) ACS routines. Because all of the logical drives and volumes are associated with the composite library, only the composite library can be defined in the SG.
For information about host software implementation tasks for IBM tape libraries, see the following resources:
•z/OS DFSMS OAM Planning, Installation, and Storage Administration Guide for Tape Libraries, SC23-6867
•IBM TS3500 Tape Library with System z Attachment A Practical Guide to Enterprise Tape Drives and TS3500 Tape Automation, SG24-6789
•IBM TS4500 R5 Tape Library Guide, SG24-8235
If your TMS is DFSMS Removable Media Manager (DFSMSrmm), see the following manuals for more information:
•DFSMSrmm Primer, SG24-5983
•z/OS DFSMSrmm Implementation and Customization Guide, SC23-6874
If this installation is the first SMS tape library at this host, additional steps are required. The full product documentation for your TMS needs to be consulted, in addition to the OAM PISA listed in the previous bullets.
Complete the following steps to define the TS7700 tape library in an existing z/OS SMStape environment:
1. Use the ISMF Library Management → Tape Library → Define panel to define the tape library as a DFSMS resource. Define the composite library and one or more distributed libraries. Library names cannot start with a “V”. In the following figures, we define one composite library that is named IBMC1 and a single distributed library that is named IBMD1.
|
Remember: Library ID is the only field that applies for the distributed libraries. All other fields can be blank or left as the default.
|
Figure 6-6 Define Composite Tape Library panel 1
Figure 6-7 shows the second panel.
Figure 6-7 Define Composite Tape Library panel 2
Figure 6-8 shows the third panel.
Figure 6-8 Define Composite Tape Library panel 3
Figure 6-9 shows defining the distributed tape library.
Figure 6-9 Define Distributed Tape Library
2. Using the Interactive Storage Management Facility (ISMF), create or update the DCs, SCs, and MCs for the TS7700. Ensure that these defined construct names are the same as those that you defined at the TS7700 MI.
3. Using ISMF, create the SGs for the TS7700. Ensure that these defined construct names are the same as those that you defined at the TS7700 MI.
The composite library must be defined in the SG. Do not define the distributed libraries in the SG.
|
Tip: At OAM address space initialization, if a distributed library is defined to an SG, the warning message CBR3017I is generated, which indicates that the distributed library is incorrectly defined to the SG.
|
4. Update the ACS routines to assign the constructs that are needed to use the TS7700 and then convert, test, and validate the ACS routines.
5. Customize your TMS to include the new volume ranges and library name. For DFSMSrmm, this involves EDGRMMxx updates for VLPOOL and LOCDEF statements, in addition to any OPENRULE and PRTITION statements needed. If you leave REJECT ANYUSE(*) in your EDGRMMxx member, you cannot use any tape volume serial numbers (VOLSERs) not previously defined to RMM.
6. Consider whether your current TMS database or control data set (CDS) has sufficient space for the added library, data set, and volume entries. For DFSMSrmm, see the topic “Creating the DFSMSrmm CDS” in DFSMSrmm Primer, SG24-5983.
7. Modify the SYS1.PARMLIB members DEVSUPxx for new categories, as described in 4.3.4, “Sharing and partitioning considerations” on page 175. The DEVSUPxx default categories should always be changed to prevent disruption of library operations. For more information, see the topic “Changing the library manager category assignments in an ATLDS” in z/OS V2R3.0 DFSMS OAM Planning, Installation, and Storage Administration Guide for Tape Libraries, SC23-6867.
8. Consider updating COMMANDxx to vary the library online at IPL if you have not already set it to come online during the definition of the library through ISMF. For more information, see 10.1.1, “DFSMS operator commands” on page 586. The OAM address space must be active for the vary to complete successfully.
9. In a Grid environment, it might be desirable to update the ALLOCxx parmlib member SYSTEM TAPELIB_PREF to balance workloads. Although the default algorithm, EQUAL, works well if the libraries under consideration have an equal number of online devices, the recommendation is to use BYDEVICES.
10. Consider the current size of the volume catalogs and the additional space that is required for the new volume and library entries. For a TS7700 with 100,000 volumes, at least 32 cylinders should be allocated. To more precisely estimate the space allocation requirements in the tape volume catalog, use the following steps:
a. Estimate the number of tape library entries and tape volume entries to be cataloged in the VOLCAT. Each tape library entry requires 320 bytes and each volume entry requires 275 bytes.
b. Divide the total number of bytes by 1024 to determine the number of kilobytes, or by 1,048,576 to determine the number of megabytes.
For more information, see “Defining Names for a Tape Volume Catalog” in z/OS DFSMS Managing Catalogs, SC23-6853.
11. Activate the new Source Control Data Set (SCDS) with the SETSMS SCDS (scdsname) command.
12. Restart the OAM address space with the F OAM,RESTART command. SMS SCDS activation initiates an OAM restart if the RESTART=YES parameter is specified in the OAM startup procedure in PROCLIB, and in that case a manual restart is not needed.
13. Define any security profiles required.
14. Vary the composite library and distributed libraries online. See 10.1.1, “DFSMS operator commands” on page 586.
15. Vary the TS7700 virtual drives online, as described in 10.1.1, “DFSMS operator commands” on page 586.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.