Tips for setting up SANs for hp OpenVMS Overview

Tips for setting up SANs for hp OpenVMS Overview
Overview
Tips for setting up
SANs for hp
OpenVMS
Fibre channel advantages
• SANs – multi-vendor, multi-OS, multi-system detached
storage – industry common components
• High performance, cabling distance & redundancy
Challenges
Manfred Kaser
• Mini-merge with ACS 8.7 (HSG80 firmware), OpenVMS
V7.3-1 plus appropriate ECOs !
HP Services
• Learning a new technology
Vortrag 3D04
© 2002 hp
DECUS Bonn 2003
• Can grow to be complex
page 1
© 2003 hp
DECUS Bonn 2003
page 2
1
Setup – Adapter cabling
Setup – HSG80 cabling with 4 SANs
Typically one adapter per SAN
One option -- each HSG80 port on a SAN with 4 SANs
• Highest redundancy/performance/capacity
• Not currently supported with HSV110s the SAN Appliance must see all the ports on the HSV
© 2003 hp
DECUS Bonn 2003
• high redundancy and high performance
page 3
© 2003 hp
DECUS Bonn 2003
page 4
2
Setup – Cabling suggestions
Setup – Device naming
• An HSG/HSV presents two device types to OpenVMS
• Label cables
– $1$GGAnnnnn: - console lun for the HSG/HSV pair
• Be consistent – very important
– $1$DGAnnnnn: - disk unit (HSG), virtual disk (HSV)
– whichever HSG/HSV cabling option you choose, use it for all
HSG/HSVs (unless some HSG DRM & some not)
• There is NO allocation class setting for an HSG/HSV
– Cable the adapters the same from system to system – first
adapters on the same SAN, etc.
– the 1 in $1$ is always 1 and cannot be changed
– recommend using the first digit or two of nnnnn to designate the
HSG/HSV pair or the site – as in $1$DGAccnnn or
$1$DGA1501:
• With GBIC style switches verify the GBIC type as they
all look very similar
• HSG80s must be set for Multibus failover and to SCSI-3
mode for OpenVMS
• Document everything and keep it current - This could
significantly reduce time for future problem resolution.
– HSG80> SET MULTIBUS_FAILOVER COPY={THIS|OTHER}
– HSG80> SET THIS SCSI_VERSION=SCSI-3
© 2003 hp
DECUS Bonn 2003
page 5
© 2003 hp
DECUS Bonn 2003
page 6
3
Setup – Device naming
Setup – Device naming
A disk unit Identifier (HSG), virtual disk OS Unit ID (HSV):
• The nnnnn in $1$xDGAnnnnn: is set via an Identifier,
not by the controller name, or disk unit number dNNN
(HSG) or virtual disk name vdNNN (HSV)
• is set on the HSV via the SAN Appliance and the HSV element
manager
• is set on the HSG via the SET dNNN IDENTIFIER=nnnnn
• is required for each unit for OpenVMS to see the disk or virtual
disk unit
• must be unique SAN/Zone-wide. It is possible to give more
than one unit the same identifier (doesn’t work well – very bad)
• recommend relating the unit, dNNN, or virtual disk, vdNNN, to
the Identifier nnnn. Reduces confusion.
• is viewed via the SAN Appliance on the HSV or
via SHOW dNNN or SHOW UNITS FULL on the HSG
– identifiers have the range 0-32767
– note – MSCP only serves devices named 0-9999
• The console lun Identifier for an HSG/HSV pair:
– is set via the SAN Appliance and the HSV Element
Manager for the HSV and via the command SET THIS
IDENTIFIER=nnnnn on the HSG
– is required for OpenVMS to see HSG/HSV disks
– must be unique among HSG/HSV pairs and disks, and be
non zero
© 2003 hp
DECUS Bonn 2003
page 7
© 2003 hp
DECUS Bonn 2003
page 8
4
WorldWide IDentifiers (WWIDs)
WorldWide IDentifiers (WWIDs)
Every adapter or controller connected to a SAN has a WWID
that is unique – the form is nnnn-nnnn-nnnn-nnnn where the
n’s are hex digits
View KGPSA adapters and WWIDs at the AlphaServer SRM
console with: “set wwidmgr -show adapter”
• KGPSA WWIDs typically begin with 1000 or 2000
LP00>>>wwidmgr -show adapter
item
adapter
WWN
[ 0] kgpsaa0.0.0.4.6
1000-0000-c920-05ab
[ 1] kgpsab0.0.0.8.6
1000-0000-c921-0ce0
the last digit (5000-nnnn-nnnn-nnnX) of a WWID for a
controller pair denotes:
•HSG
•0
•1
•2
•3
•4
HSV
0 = the controller pair
9 = Port 1 bottom controller
8 = Port 2 bottom controller
D = Port 1 top controller
C = Port 2 top controller
3
1
4
• As of OpenVMS V7.2-2 the WWID of a KGPSA is viewable with the
SHOW DEVICE/FULL command on the adapter device, ie:
2
$ SHOW DEVICE/FULL FGA0:
• Can also be viewed from the HSG/HSVs (more later)
• All WWIDs in the SAN can be viewed from the switches
• Document all adapter and controller WWIDs as they will be
needed for troubleshooting
•Can also be viewed from OpenVMS (more later)
DECUS Bonn 2003
Next Topo
FABRIC
FABRIC
[9999] All of the above
View HSV WWIDs from the SAN Appliance, and HSG WWIDs
with SHOW THIS/OTHER
© 2003 hp
Cur. Topo
FABRIC
FABRIC
page 9
© 2003 hp
DECUS Bonn 2003
page 10
5
Paths / Connections
Path / Connection status
A path or connection can be considered
A path or connection is defined by the two end points or
WWIDs (the adapter & the HSG/HSV port)
• Good or Bad meaning it is functional or not
• Valid or Invalid meaning it exists or not
• Connected (active) or not connected
– it is connected if the path’s HSG/HSV port is on the
controller currently serving the disk
– it is not connected if
disk unit would have to
fail-over to the other
HSG/HSV to be accessed
on the path
A B
1
4 paths per disk unit
on each system
© 2003 hp
3
4
If the paths A-1 (adapter A to HSG
port 1) & B-2 are connected then
the paths on A-3 & B-4 are not
connected, and visa-versa
8 paths per disk unit
on each system
*HSVs may only be connected to 2
SANs
DECUS Bonn 2003
2
page 11
© 2003 hp
DECUS Bonn 2003
page 12
6
AlphaServer console access to SAN Disks
Using wwidmgr
• The wwidmgr users guide is on the AlphaServer console
CD in the doc directory as wwidmgr.pdf and wwidmgr.ps
• Due to the long WWIDs fibre channel devices cannot be
incorporated into a console device name
• On AlphaServer 8x00, 4x00, or 1200 systems, the console
must be in diagnostic mode before using wwidmgr.
• Fibre channel devices must have an association defined
between the device name and WWID before they can be
accessed from the console
– P00>>> set mode diag
• Since the context is not the same as when the system
boots an INIT may be necessary before entering diag
mode.
– only FC based system (with shadowset members) and dump disks
require console access
• Use the utility wwidmgr to manage the associations
• After using wwidmgr the system must be re-initialized
before booting or to exit wwidmgr mode.
• 8 environment variables to store association info
– N1-N4 are used for the WWIDs to access
– wwid0-wwid3 are used for device names
– only 4 devices and 4 paths can be assigned
– P00>>> init
• Wwidmgr commands have the syntax:
– P00>>> wwidmgr command [–option [parameter]]
© 2003 hp
DECUS Bonn 2003
page 13
© 2003 hp
DECUS Bonn 2003
page 14
7
OpenVMS & FibreChannel paths
Using wwidmgr
Remember – a path is defined by two endpoints, such as two
WWIDs. OpenVMS uses the adapter name and the WWID
of the HSG/HSV port. For example:
Remember: HSG/HSV console lun & disk units must be
assigned identifiers before OpenVMS can access them.
These can be used with wwidmgr, too.
$ show device /fu $1$DGA706
Disk $1$DGA706: (MUN01), device type HSG80, is online, file-oriented device,shareable,
device has multiple I/O paths, served to cluster via MSCP Server, error logging is enabled.
displaying FC devices
• To poll each FC adapter and display devices found
5
. . . I/O paths to device
Path MSCP (MUNAXP), primary path.
Error count
Path PGA0.5000-1FE1-0004-EDB1
Error count
Path PGA0.5000-1FE1-0004-EDB3
Error count
Path PGA0.5000-1FE1-0004-EDB4
Error count
Path PGA0.5000-1FE1-0004-EDB2
Error count
– wwidmgr –show wwid
– wwidmgr –show wwid –full
!displays all paths
– (ev:none) means no environment variable is set for it
• To display a specific device
– wwidmgr –show wwid –full –udid <identifier>
0
Operations completed
(MUN01).
0
Operations completed
(MUN01), current path.
0
Operations completed
(MUN01).
0
Operations completed
(MUN01).
0
Operations completed
0
0
52304
0
0
ƒ The primary path is the first one found by OpenVMS at I/O
configuration.
ƒ The current path is the one in use now.
© 2003 hp
DECUS Bonn 2003
page 15
© 2003 hp
DECUS Bonn 2003
page 16
8
OpenVMS & Changing paths
OpenVMS & FibreChannel paths
A path can be set manually via
OpenVMS Multipath devices have an UCB for every path.
These can be viewed with ANALYZE/SYSTEM:
$ SET DEVICE/PATH=device-path/SWITCH
A path can be changed automatically by a mount
verification.
SDA> SHOW DEVICE $1$DGA#####
The current path (or UCB) is the one to be used for the next
I/O for the device.
• Changing the current path manually has no effect on
the HSG/HSV controllers or the rest of the OpenVMS
Cluster until the next I/O is issued.
– bad another will be chosen via a mount verification
– not connected, then the HSG/HSVs will attempt to fail the disk
over so this path will be connected – if the disk is mounted
elsewhere in the cluster paths will be disrupted and new paths
chosen via mount verification
DECUS Bonn 2003
– Before OpenVMS 7.2-2 mount verification attempts the
primary path first
Starting with OpenVMS V7.3-1, mounting a multipath disk
can result in a path change.
• When an I/O is attempted, if the current path is:
© 2003 hp
– Starting with OpenVMS V7.2-2 (or earlier with VMS721_SYSV0100 or VMS721H1_SYS-V0600) mount verification
attempts the current path first.
page 17
When a path is changed an OPCOM message is generated which
includes whether the change was manual or automatic. Note that path
changes may result in Mount Verifications, too.
© 2003 hp
DECUS Bonn 2003
page 18
9
OpenVMS & FibreChannel paths
OpenVMS & Path Balancing
The current path of FC devices can be viewed via
$ SHOW DEVICE/MULTIPATH
$ show device/mult
Device
Name
$1$DGA1:
(LILDOG)
$1$DGA3:
(LILDOG)
$1$DGA45:
(LILDOG)
$1$DGA46:
(LILDOG)
Device
Name
$1$GGA99:
$
$ sho dev dg/mult
Device
Name
$1$DGA1:
(LILDOG)
$1$DGA3:
(LILDOG)
$1$DGA45:
(LILDOG)
$1$DGA46:
(LILDOG)
© 2003 hp
Device
Status
Online
Mounted alloc
Online
Mounted
Device
Status
Online
Error Paths Current
Count Avl/Tot
path
0
4/ 4 PGA0.5000-1FE1-0005-35E3
2
4/ 4 PGA0.5000-1FE1-0005-35E3
2
4/ 4 PGA0.5000-1FE1-0005-35E3
7
4/ 4 PGB0.5000-1FE1-0005-35E2
Error Paths Current
Count Avl/Tot
path
0
4/ 4 PGB0.5000-1FE1-0005-35E2
Device
Status
Online
Mounted alloc
Online
Mounted
Error Paths Current
Count Avl/Tot
path
0
3/ 4 PGA0.5000-1FE1-0005-35E3
2
4/ 4 PGA0.5000-1FE1-0005-35E3
2
4/ 4 PGA0.5000-1FE1-0005-35E3
8
4/ 4 PGB0.5000-1FE1-0005-35E2
DECUS Bonn 2003
page 19
Path balancing is the selection of current paths so
all adapters and controller ports available are
utilized.
Recall that a path is defined by two end points,
adapter and controller port. Once set, the path
used only changes if a mount verification occurs,
or there is manual intervention.
Observation is most sites choose to do path
balancing. But, is it always needed?
• With fibre channel’s high performance characteristics,
bottlenecks are less common. So, path balancing may not
improve performance.
• However, it does enhance failure detection.
© 2003 hp
DECUS Bonn 2003
page 20
10
OpenVMS & Path Balancing
OpenVMS & Path Balancing
Prior to OpenVMS V7.3-1
The default path selection behavior was the current path would be
the primary path.
Path balancing was a manual process and a system management
responsibility.
• Path balancing could only be done from the OpenVMS
system. Since OpenVMS did not distinguish whether or not a
path was connected, HSG preferred path settings were
meaningless.
Question – If paths are manually balanced on all cluster nodes
and one node was rebooted, by default, what happened to the
path settings?
Answer – The primary path was chosen as the current path.
When the disks were mounted some units failed-over on
the HSG/HSVs, resulting in path changes on other
nodes.
© 2003 hp
DECUS Bonn 2003
page 21
Prior to OpenVMS V7.3-1
Recommended procedure to balance paths:
•
Create a command procedure for all SET DEVICE/PATH
commands, so that all adapters and controller ports are
used.
•
Execute the procedure prior to mounting disks.
•
If the balance of paths get out of synch. due to some
event, Execute the procedure again. Recall, doing this
will result in mount verification messages.
•
If possible use the same procedure on all applicable
cluster nodes. This reduces unit fail-over on controllers,
but does requires consistent cabling from node to node.
© 2003 hp
DECUS Bonn 2003
page 22
11
OpenVMS & Path Balancing
OpenVMS & Path Balancing
New with OpenVMS V7.3-1
Automatic path balancing for multipath disk and tape devices
occurs during certain events.
• When a new path to a multipath device is configured, the direct
path with the fewest devices is chosen automatically as the
current path.
• The MOUNT utility can, for a disk, trigger an automatic path
selection search biased toward the connected path presently
used the least as a current path.
• Benefits include:
–
–
–
–
–
© 2003 hp
Minimizing the disruption on other hosts in the cluster.
Tends to preserve any manual load balancing done on other cluster nodes.
Enables the use of HSG/V capabilities for initial default distribution of
devices between the two controllers.
Tends to balance the use of available paths from this host to the disk
devices.
Prefers direct paths over MSCP served paths.
DECUS Bonn 2003
page 23
New with OpenVMS V7.3-1
Path selection during the MOUNT of a multipath disk, is as follows:
1. If the current path is a direct path with a connected status it is used.
2. If the current path is an MSCP path selected manually by a path
switch command it is used.
3. All direct paths are checked. They are considered in order of
increasing use as a current path for other devices. The first connected
path found is selected as the current path. If none were connected,
the first path considered is used, and an attempt is made to fail over
the device on the HSG/HSV pair to make the path connected.
4. The MSCP served path is tried.
5. If the result was a current path change, an automatic path switch is
performed and an OPCOM message is issued.
The MOUNT utility can trigger this path selection algorithm a number of
times until a working path is found. The number of retries depends on
the elapsed time and the qualifiers specified with the command.
© 2003 hp
DECUS Bonn 2003
page 24
12
OpenVMS & Path Balancing
OpenVMS & Path Balancing
New with OpenVMS V7.3-1
Automatic path selection algorithms give meaning to
preferring paths on the HSG/HSVcontrollers:
•
New with OpenVMS V7.3-1
Recommended procedure to balance paths:
Set the preferred HSG controller (top/bottom) to serve each
unit. Note - only when both controllers are rebooted does this
have an effect.
SET unit PREFERRED_PATH={THIS|OTHER}_CONTROLLER
•
HSVs have preferred path and preferred fail-back path
selection via the SAN Appliance and the HSV Element
Manager:
– Setting preferred paths only on HSV has the same effect as on
HSGs.
– Setting preferred fail-back effects recovery when a member of an
HSV pair fails and later recovers. Note – when the controller
recovers and makes the fail-back settings, path changes (mount
verifications) result on the OpenVMS systems.
© 2003 hp
DECUS Bonn 2003
page 25
•
Set the preferred HSG/HSV controller (top/bottom) to serve
each unit. On an HSV, optionally set the fail-back paths.
•
Create a command procedure for all SET DEVICE/PATH
commands, so that all adapters and controller ports are used.
Match the settings to the controller preferred path settings.
•
Execute the procedure after disks are mounted as needed.
Recall, doing this will result in mount verification messages.
•
If possible use the same procedure on all applicable cluster
nodes. This reduces unit fail-over on controllers, but does
requires consistent cabling from node to node.
© 2003 hp
DECUS Bonn 2003
page 26
13
OpenVMS & FibreChannel paths
OpenVMS & Adding FibreChannel Devices
FC utilizes SCSI protocols. So, devices appear as local
devices to OpenVMS.
New with OpenVMS V7.3-1
When a unit is added on an HSG/HSV, and an (OS) identifier
assigned, it does not automatically appear on OpenVMS.
• An autoconfigure is required for a new unit to appear on
an OpenVMS node. Recommended to do this clusterwide.
Switching between MSCP served and direct paths for
multipath disks. Great news for multi-site OpenVMS
Clusters with extended SANs
•
$ MCR SYSMAN
SYSMAN> SET ENVIRONMENT/CLUSTER
SYSMAN> IO AUTOCONFIGURE
Prior to OpenVMS V7.3-1 (or in V7.3-1 with MSCP serving
disabled), in an OpenVMS Cluster, it was very important
to autoconfigure on all nodes before using a new device.
• Disabling MSCP serving (if not needed) of SAN devices
was also recommended prior to OpenVMS V7.3-1
© 2003 hp
DECUS Bonn 2003
page 27
•
© 2003 hp
If a multipath disk is enabled for MSCP serving and its
direct paths fail its current path will change to an MSCP
path. If one or more direct paths are then restored it will
fail back to a direct path.
Use the MPDEV_REMOTE system parameter to enable
this capability.
To manually switch the current path to an MSCP served
path, use
SET DEVICE device-name/SWITCH/PATH=MSCP
DECUS Bonn 2003
page 28
14
Two-site Host Based Volume Shadowing
Firmware and ECOs
With OpenVMS V7.2-2, HBVS over extended SANs is supported.
It is also available for OpenVMS V7.2-1 and V7.2-1H1 via a Shadowing
ECO, but the “Enhanced Volume Shadowing functionality” must be
chosen during installation.
– Adds new member timeout qualifier
to force manual intervention in the
event the SAN inter-site link is lost,
but not the SCS link. This avoids the
possibility of split shadow sets.
Not necessary with OpenVMS V7.3-1
if disks are MSCP served.
– New /SITE and /READ_COST
qualifiers were added since all the
disks appear local.
– See the ECO release notes and/or the current Volume Shadowing for
OpenVMS manual, section 4.7 – available on the web at
http://www.openvms.compaq.com:8000/index.html
© 2003 hp
DECUS Bonn 2003
Fibre Channel is a relatively new and evolving technology. So,
staying current on firmware levels (for controllers, switches and
adapters) and related OpenVMS ECOs is important. Check
firmware versions on:
• HSGs via SHOW THIS or SHOW OTHER, current is ACS V8.7-1, or
ACS V8.6x-12. HSVs via the SAN Appliance, current is VCS V2.0.
• Switches with the telnet console “version” command.
• Adapters via ANALYZE/SYSTEM - checking FG devices. Firmware
is updated with the
SDA> fc show device fga0
FGA0: operational firmware revision DS3.03X2
console. See the
fc show device fgb0
current Alpha console SDA>
FGB0: operational firmware revision DS3.03X2
firmware CD.
– On OpenVMS check for FIBRE-SCSI, SYS, SHADOWING and
UPDATE ECOs with PRODUCT SHOW HISTORY. Current
OpenVMS ECOs are available at
http://ftp1.support.compaq.com/public/vms/
page 29
© 2003 hp
DECUS Bonn 2003
page 30
15
HSG80 Connections
HSG80 Connections
HSG80TOP_>show connection
Connection
Name
Operating system
On an HSG a path is called a Connection. It is defined by
the adapter WWID and the HSG port.
Connections are automatically created when a system
accesses the controller for the first time
• “wwidmgr –show wwid”, booting OpenVMS, etc. should create any connections as needed.
• Connections can be created manually – but that is not
generally recommended
HSG80s have a limit of 96 defined connections. It is
recommended to delete unneeded or inactive connections.
Controller
Port
Address
Status
Unit
Offset
!NEWCON13
VMS
OTHER
HOST_ID=2000-0000-C922-4C03
1
011200
OL other
0
ADAPTER_ID=1000-0000-C922-4C03
!NEWCON14
VMS
OTHER
HOST_ID=2000-0000-C922-9C4C
2
011600
OL other
0
ADAPTER_ID=1000-0000-C922-9C4C
!NEWCON15
VMS
THIS
HOST_ID=2000-0000-C922-4C03
1
011200
OL this
0
ADAPTER_ID=1000-0000-C922-4C03
!NEWCON16
VMS
THIS
HOST_ID=2000-0000-C922-9C4C
2
011600
OL this
0
ADAPTER_ID=1000-0000-C922-9C4C
Very important the “operating system” is set correctly – the
default is WINNT or the setting of the last connection set.
HSG80> SET connection-name OPERATING_SYSTEM=VMS
Recommend re-naming connections (max. of 9 characters)
> DELETE connection-name
HSG80> RENAME !NEWCON13 LildogA
Note a connection cannot be deleted if it has access
paths currently enabled
HSG80s do not have names - recommend setting the prompt
HSG80> SET THIS PROMPT=“HSG80_99Top> “
© 2003 hp
DECUS Bonn 2003
page 31
© 2003 hp
DECUS Bonn 2003
page 32
16
HSV Connections
Switch settings
On an HSV a path is called a Connection. It is defined by the Host
adapter WWID and the HSV port.
Host adapter access is managed/configured using the HSV Element
Manager from the SAN Appliance. Host port entries are automatically
created when a system accesses the controller for the first time
• “wwidmgr –show wwid”, booting OpenVMS, etc. - should create
any connections as needed.
• Hosts and host ports can be created manually – but that is not
generally recommended
A host must be add/configured on the HSV before it can access any
units. First you add a host, which includes one port (adapter). It is
very important to set the operating system properly in the pull-down
menu. Additional ports can then be added to a host.
It is recommended to delete unneeded or inactive hosts.
> DELETE connection-name
Note a connection cannot be deleted if it has access paths currently
enabled
© 2003 hp
DECUS Bonn 2003
page 33
Setting the switch domain_id can aid in troubleshooting.
To set the FC switch host_id at the telnet console, the switch has to be
disabled first:
admin> switchDisable
admin> flashset
MAIN09:admin> switchShow
switchName:
MAIN09
domain_id [0]: 9 !Must be unique in the fabric and
switchType:
2.4
:
switchState:
Online
!if possible
.
switchRole:
Principal
switchDomain:
9
admin>
switchEnable
switchId:
fffc01
switchWwn:
10:00:00:60:69:10:5c:de
port 0: sw Online
F-Port 50:00:1f:e1:00:00:f7:03
port 1: sw Online
F-Port 50:00:1f:e1:00:00:f7:01
port 2: sw No_Sync
port 3: sw Online
F-Port 10:00:00:00:c9:24:36:2b
port 4: sw Online
F-Port 50:00:1f:e1:00:05:63:d4
port 5: sw Online
F-Port 50:00:1f:e1:00:05:d8:81
port 6: sw Online
F-Port 10:00:00:00:c9:23:81:e1
port 7: id Online
F-Port 50:00:1f:e1:00:05:d8:83
port 8: id Online
F-Port 10:00:00:00:c9:22:47:5f
port 9: id Online
F-Port 50:00:1f:e1:00:05:63:d1
port 10: id No_Light
port 11: sw Online
F-Port 10:00:00:00:c9:22:79:c3
port 12: sw Online
F-Port 10:00:00:00:c9:23:01:4d
port 13: -- No_Module
port 14: -- No_Module
port 15: -- No_Module
© 2003 hp
DECUS Bonn 2003
unique for the site.
page 34
17
HSG/HSV Console Access
Additional Resources
The only fully functional HSG management access is via the console port.
• Scripting access for most commands is available via the
SANWorks Command Scripter product. It utilizes the SWCC agent
to access the controller.
• HSZTerm (Set Host/SCSI) works via the GGA device. It is officially
not supported. Using while a SWCC agent is running can crash
the system. Provides access to all commands and utilities except
VTDPY.
Section 6.6 & 6.7 and Chapter 7 of the Guidelines for OpenVMS
Cluster Configurations manual contain excellent OpenVMS Fibre
Channel configuration information.
• The OpenVMS Documentation set is available on the web at
A good console manager is the best solution for an HSG providing:
• Console access (via the web and command line/terminal)
• Scripting access
• Console output logging. This is very important as this will capture
all error output from the controller.
DECUS Bonn 2003
For details on the fibre channel related enhancements in OpenVMS
V7.3-1, see sections 4.9 and 4.16 of the OpenVMS Alpha Version
7.3--1 New Features and Documentation Overview
For the latest white papers, including “Fibre Channel in a DisasterTolerant OpenVMS Cluster System”, and information about fibre
channel on OpenVMS visit the web page
http://www.openvms.compaq.com/openvms/fibre/
Full management access for the HSV is provided via the SAN Appliance.
Both GUI (SANWorks Element Manager) and character-cell (Enterprise
Storage System Scripting Utility (SSSU)) interfaces are available from
the Appliance.
© 2003 hp
http://www.openvms.compaq.com:8000/index.html
page 35
For documentation on the StorageWorks enterprise virtual array
(HSV) visit the web page
http://www.compaq.com/products/storageworks/enterprise/documentation.html
© 2003 hp
DECUS Bonn 2003
page 36
18
Summary
Cabling consistency for systems & controllers is key
Name HSG80 connections, set the O/S, & set switch host_ids
Maintain SAN configuration documentation in common formats that
can be shared via e-mail
Keep current on firmware (controller, adapter & switch) and ECOs.
Do not mix versions of firmware.
Questions?
If path balancing, create a command procedure. Prior to OpenVMS
V7.3-1 execute it before mounting disks during system startup.
Path changes can cause mount verifications and mount verifications
can cause path changes.
Remember to IO AUTOCONFIGURE on all Cluster FC nodes when
adding devices/units.
Console management is recommended for HSG80s.
© 2003 hp
DECUS Bonn 2003
page 37
© 2003 hp
DECUS Bonn 2003
page 38
19
20
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising