nsc-uw7/ 0000755 0000175 0000001 00000000000 07255423005 0013215 5 ustar 00cfbuild other 0001776 0002004 nsc-uw7/eng/ 0000755 0000175 0000001 00000000000 07255423005 0013766 5 ustar 00cfbuild other 0001776 0002004 nsc-uw7/eng/compaq/ 0000755 0000175 0000001 00000000000 07255423146 0015254 5 ustar 00cfbuild other 0001776 0002004 nsc-uw7/eng/compaq/docs/ 0000755 0000175 0000001 00000000000 07255423077 0016207 5 ustar 00cfbuild other 0001776 0002004 nsc-uw7/eng/compaq/docs/Cim_user.htm 0000644 0000175 0000001 00000012220 07255423077 0020464 0 ustar 00cfbuild other 0001776 0002004
The Compaq Insight Manager Agents provide SNMP access to useful system information. These agents are installed with the installation script by default.
IMPORTANT: This software can be installed on the 7.1.1 or later versions of SCO UnixWare 7 NonStop Clusters. It upgrades the Compaq Insight Manager agents or installs an initial version of Compaq Insight Manager Agents for SCO UnixWare 7 NonStop Clusters version 7.1.1. |
The agent packages installed are as follows:
shutdown -i6 -g0 -y
Wait while the system configures hardware and loads configuration files.
onall /etc/init.d/cpqsnmp stop_for
onall /etc/init.d/snmp_node stop
onall /etc/init.d/cpqsnmp start_for
/etc/init.d/cpqsnmp start | stop
/etc/init.d/cpqsnmp start_for | stop_for node
dlpid: InfoInterface: Unable to find DLPI Interface (DEVICE) in internal tableThis message is harmless, but can be removed by stopping and starting the cmanic daemon. Use the following steps:
onall /etc/init.d/cmanic stop
onall /etc/init.d/cmanic start
onall /etc/init.d/cmanic start
The nscvu(1M) utility reports information about the configuration of a newly installed cluster. It provides cluster-wide information as well as node-specific information, volume information, and external storage information. This utility is installed by default when you use the installation script. For information about nscvu, refer to the nscvu(1M) manual page, which is viewable with the man command or through the SCOHelp online documentation.
The manual page is also included on this CD, as follows:
Note that hypertext links to additional manual pages do not work while you view the manual page on this CD-ROM. You must view the installed manual page using SCOHelp to access the additional manual pages.
The nscvu command:
nscvu [-f [-x]] [filename]
The nscvu utility reports the system configuration of a newly installed cluster. This utility gathers information from various sources on the cluster, correlates that information, and produces a report which details the configuration and notes any abnormal conditions encountered during the process. The utility will optionally perform filesystem failover tests to validate that configured filesystems remain accessible in the event of a node failure. By default, only non-disruptive tests are performed.
filename
The output of this utility is a report in six sections:
In the report output, a value of N/A in a field indicates that information is not available. An asterisk in front of a line in the report indicates an abnormal conditions involving data on that line. A question mark (?) in front of a line indicates that error status for the line could not be determined. This is usually because information for one or more fields was not available.
The cluster-wide section of the report contains the following information:
The cluster-wide parameters section of the report is similar to the following:
ProLiant NonStop Clusters Verification Utility Version 1.1 (nscvu) Wed Dec 1 14:42:05 1999 Cluster name: your_cluster, CVIP: XXX.XXX.XXX.XXX, 2 nodes SCO UnixWare 7.1.1 -- NonStop Clusters, version: E00 08/11 04:17 ServerNet interconnection, status: operational, SHIP 1.5 Rev E Summary status: operational
In an actual report, your_cluster is your cluster name and the XXX.XXX.XXX.XXXs represent the IP address for the CVIP associated with your cluster name.
The second section of the report contains information related to each specific node in the cluster:
The node-specific parameters section of the report is similar to the following:
Node Information Node Stat IP Model Mem ServerNet Processor Cache 1 UP XX.XX.XX.XXX PL1600 64 operational 300 MHz P-II 512K 300 MHz P-II 512K 2 UP XX.XX.XX.XXX PL1600 64 operational 300 MHz P-II 512K
In the actual output, the XX.XX.XX.XXXs are replaced by the appropriate IP addresses.
External storage information is associated with the controlling node pair. This section of the report lists each disk array connected to a two-node pair, the SCSI devices provided, and the controlling nodes. For each disk, information listed includes the SCSI ID of each logical disk, the filesystem mountpoint (if any), the capacity, the disk status, whether the device is configured for failover, the mirroring status, and comments on any unusual conditions.
For each Fibre Channel storage box, the following information is displayed:
The Fibre Channel external storage information section of the report is similar to the following:
External Storage FCAL Array: S/N: D908HUQ10078 Controlled by: 1(0) 2(0) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b0t0d0 / 8673 Online YES RAID-1 Veritas rootdg 1 c2b0t1d0 - 8673 Online NO Not Mirrored 2 c2b0t2d0 - 4091 Online NO Not Mirrored 3 c2b0t3d0 - 4091 Online NO Not Mirrored
This section of the report contains the failover file systems and their status. If the storage box is not associated with any cluster nodes, this section is not included in the report. If any devices are incorrectly configured for failover, their status is included. The following entries appear in this section:
The Fibre Channel file system failover section of the report is similar to the following:
Failover Filesystems Nodes Device Type Valid Matched Comments ? 1, 2 c2b0t1* Fs Yes Yes Veritas 1, 2 c2b0t2* Fs Yes NO MATCH
In this example, all devices for nodes 1 or 2 having a SCSI ID beginning with c2b0t1 are configured for failover. Some devices did actually match the pattern specified; however, the devices are managed by Veritas and that subsystem is responsible for managing failover operation. The second rule did not match any devices available to the system.
This section includes information about Veritas disk groups and the mirrored volumes they contain. This section is not included in the report if no mirrors exist. For each disk group, the following information is displayed:
Additionally, a list of all dissociated plexes (disk plexes which are not part of a volume) is presented.
The mirrored volume information section of the report is similar to the following:
Veritas Disk Group: rootdg Volume Fsys P1-Media P2-Media CNM Status Comments standvol - n1c0b0t0d0s0 - - Not mirrored swapvol - n1c0b0t0d0s0 - - Not mirrored rootvol - n1c0b0t0d0s0 n2c0b0t0d0s0 YES Mirrored Dissociated plexes for rootdg Plex Media Comments rootvol-01 n1c0b0t0d0s0 osuvol-01 n1c2b0t0d0s0
This section includes key information about the cluster status. The section includes information such as:
A complete list of Analysis Summary messages and their meanings is presented below. The analysis summary section of the report is similar to the following:
Analysis Summary * SNMP is not installed on the cluster. * No ServerNet switches detected. Acceptable. * No external storage devices detected. * No failover information detected.
ProLiant NonStop Clusters Verification Utility Version 1.1 (nscvu) Wed Dec 1 11:02:36 1999 Cluster name: your_cluster, CVIP: XX.XX.XX.XX, 2 nodes SCO UnixWare 7.1.0 NonStop Clusters, version: E00 08/11 04:17 ServerNet interconnection, status: operational, SHIP 1.5 Rev E Summary status: operational Node Information Node Stat IP Model Mem ServerNet Processor Cache 1 UP XX.XX.XX.XX PL850R 64 operational 200 MHz P-Pro 256K 200 MHz P-Pro 256K 2 UP XX.XX.XX.XX PL850R 96 operational 200 MHz P-Pro 256K 200 MHz P-Pro 256K Veritas Disk Group: rootdg Volume Fsys P1-Media P2-Media CNM Status Comments standvol - n1c0b0t0d0s0 - - Not mirrored swapvol - n1c0b0t0d0s0 - - Not mirrored rootvol / n1c0b0t0d0s0 n2c0b0t0d0s0 YES Mirrored
Dissociated plexes for rootdg Plex Media Comments rootvol-01 n1c0b0t0d0s0 Analysis Summary * No ServerNet switches detected. Acceptable. * No external storage devices detected. * No failover information detected. * Disk group "rootdg" contains 1 disassociated plex(es).
ProLiant NonStop Clusters Verification Utility Version 1.1 (nscvu) Wed Dec 1 11:06:47 1999 ***** NOTE: SNMP is not installed! ***** Cluster name: your_cluster, CVIP: XX.XX.XX.XX, 6 nodes SCO UnixWare 7.1.0 -- NonStop Clusters, version E00 08/11 04:17 ServerNet interconnection, status: operational, SHIP 1.5 Rev E ? Summary status: indeterminate Node Information Node Stat IP Model Mem ServerNet Processor Cache ? 1 UP XX.XX.XX.XX N/A N/A operational N/A N/A ? N/A N/A ? 2 UP XX.XX.XX.XX N/A N/A operational N/A N/A ? N/A N/A ? 3 UP XX.XX.XX.XX N/A N/A operational N/A N/A ? N/A N/A ? 4 UP XX.XX.XX.XX N/A N/A operational N/A N/A ? N/A N/A ? 5 UP XX.XX.XX.XX N/A N/A operational N/A N/A ? 6 UP XX.XX.XX.XX N/A N/A operational N/A N/A External Storage FCAL Array: S/N: D908HUQ10078 Controlled by: 1(0) 2(0) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b0t0d0 / 8673 Online YES RAID-1 1 c2b0t1d0 - 8673 Online NO Not Mirrored 2 c2b0t2d0 - 4091 Online NO Not Mirrored 3 c2b0t3d0 - 4091 Online NO Not Mirrored 4 c2b0t4d0 - 8673 Online NO Not Mirrored 5 c2b0t5d0 - 8673 Online NO Not Mirrored FCAL Array: Bus 1 S/N: D908HUQ10047 Controlled by: 3(0) 4(1) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b2t0d0 - 4091 Online NO Not Mirrored 1 c2b2t1d0 - 4091 Online NO Not Mirrored 2 c2b2t2d0 - 4091 Online NO Not Mirrored 3 c2b2t3d0 - 4091 Online NO Not Mirrored 4 c2b2t4d0 - 4091 Online NO Not Mirrored 5 c2b2t5d0 - 4091 Online NO Not Mirrored FCAL Array: Bus 0 S/N: D908HUQ10046 Controlled by: 3(1) 4(2) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b0t0d0 - 4091 Online NO Not Mirrored 1 c2b0t1d0 - 4091 Online NO Not Mirrored 2 c2b0t2d0 - 4091 Online NO Not Mirrored 3 c2b0t3d0 - 4091 Online NO Not Mirrored 4 c2b0t4d0 - 4091 Online NO Not Mirrored 5 c2b0t5d0 - 4091 Online NO Not Mirrored FCAL Array: array 0 S/N: D908HUQ10119 Controlled by: 3(2) 4(0) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b1t0d0 - 17363 Online NO RAID-1 1 c2b1t1d0 - 17363 Online NO RAID-1 2 c2b1t2d0 - 17363 Online NO RAID-1 3 c2b1t3d0 - 17363 Online NO RAID-1 FCAL Array: S/N: D908HUQ10102 Controlled by: 5(0) 6(0) Disk SCSI-ID Fsys Size Status FDsk MStatus Comments 0 c2b0t0d0 - 17363 Online NO Not Mirrored 1 c2b0t1d0 - 2004 Online NO Not Mirrored 2 c2b0t2d0 - 8673 Online NO Not Mirrored 3 c2b0t3d0 - 8673 Online NO Not Mirrored 4 c2b0t4d0 - 8673 Online NO Not Mirrored 5 c2b0t5d0 - 8673 Online NO Not Mirrored Failover Filesystems Nodes Device Type Valid Matched Comments 1, 2 c2b0t1* Auto Yes Yes 1, 2 c3b0t2* Manual Yes NO MATCH 1, 2 c3b0t3* Manual Yes NO MATCH Analysis Summary * SNMP is not installed on the cluster. * No Veritas mirrored volumes have been defined.
SNMP is not installed on the cluster.
Required SNMP utilities were not found in the expected location on this cluster. The SNMP agents are probably
not installed. Some information obtained via SNMP will not be available.
Node n is in state: state.
The specified node is in some state other than "UP". For optimal cluster operation, all nodes should
be UP.
No IP address available for node n.
The network configuration information provided by ifconfig(1M) did
not include an IP address for the specified node. For proper operation, all nodes should have an IP address defined.
Since the address is missing, it will not be possible to retrieve information about the node from SNMP.
Cannot get model for node n. Bad SNMP data.
An SNMP query for model information for the given node did not return any useful information. One or more SNMP
agent daemons on the node may be down.
Cannot get memory for node n. Bad SNMP data.
An SNMP query for memory information for the given node did not return any useful information. One or more
SNMP agent daemons on the node may be down.
No failover information detected for non-root filesystems.
Any disks not under Veritas control must be configured in /etc/cluster/fdevtab
if they are to be remounted in case their controlling node fails.
One or more Filesystem Failover rules match Veritas managed disks.
Disks managed by Veritas should not have their physical disks mentioned in /etc/cluster/fdevtab.
Veritas is responsible for making the storage available again after a node failure.
Cannot perform failover tests. No available filesystems.
Either no filesystems were configured for failover or the only configured filesystems are controlled by nodes
that are part of a cross-node mirror. A separate option to nscvu must
be specified to include nodes containing cross-node mirrored filesystems in failover tests.
Filesystem failover tests have all passed.
The requested filesystem failover tests have all completed successfully. Output from those tests is stored
in a separate file.
Node number (n1) does not match node for device (n2).
Aliases for failover devices may specify both a node number and a node-encoded device name. In all cases, the
node number should match the node number contained within the node-encoded device name.
Aliased devices for different nodes do not match.
In a failover rule for disks in an FCAL storage system, the device name, exclusive of the node ID, should be
identical for all aliases.
Device dev defined twice for node n.
In a failover rule, a device on a given node should be mentioned only once.
File filename for node n contains invalid data after failover.
The temporary file created on the filesystem expected to be present after the failure of the specified node
was found but did not contain the expected data.
File filename for node n not found after failover.
The temporary file created on the filesystem configured for failover was missing after the specified node went
down.
Unable to validate ServerNet links to other nodes from node n.
The spam(1M) utility was unable to gather data about the active ServerNet
paths from the specified node.
Unable to gather ServerNet link data to other nodes from node n.
The spam(1M) utility was unable to gather data regarding the operational
status of the X and Y paths from the specified node to the other nodes in the cluster.
ServerNet status indeterminate since node n is not UP.
Complete analysis of ServerNet connections requires all nodes to be UP and operational. If the node is not
UP, its ServerNet paths cannot be verified.
Differing SHIP card revisions detected [revisionList].
All cluster nodes are expected to contain the same revision of SHIP card. Differing versions may still be compatible,
but they should be checked.
No ServerNet switches detected. Acceptable.
In a two-node cluster, ServerNet switches are not required to connect the nodes. None were detected during
analysis. If switches really are present, the problem should be investigated.
X-side ServerNet switch is not operational. Correct, then rerun nscvu.
The cluster uses ServerNet switches for interconnection and the ServerNet switch for the X side could not be
reached from any node.
Y-side ServerNet switch is not operational. Correct, then rerun nscvu.
The cluster uses ServerNet switches for interconnection and the ServerNet switch for the Y side could not be
reached from any node.
Node n cannot communicate with X-side ServerNet switch. Correct, then rerun nscvu.
Both ServerNet switches appear to be operational; however, the X-side switch could not be reached from the
specified node.
Node n cannot communicate with Y-side ServerNet switch. Correct, then rerun nscvu.
Both ServerNet switches appear to be operational; however, the Y-side switch could not be reached from the
specified node.
Node n1 cannot reach node n2 via X path.
Both ServerNet switches are operational and the specified node can reach the X-side switch; however, it is
unable to reach the other node using the X path.
Node n1 cannot reach node n2 via Y path.
Both ServerNet switches are operational and the specified node can reach the Y-side switch; however, it is
unable to reach the other node using the Y path.
FCAL Array name S/N: serialnum is only controlled by one node (n).
FCAL arrays may be controlled by either one or two nodes. For NonStop Clusters, it is always recommended to
use two nodes in order to avoid a single point of failure.
Array S/N serialnum controlled by more nodes than defined for failover rule.
A given filesystem failover rule defines a list of aliases for a single filesystem. A disk contained in an
FCAL array will have a different node-encoded name for each controlling node. The failover rule should specify
each of these names to achieve the maximum coverage that the hardware will support.
FCAL Array S/N serialnum disk scsiID matched by multiple failover rules.
A given disk should only match a single failover rule. Since failover rules may include wildcards, it is possible
to inadvertently define multiple rules which would apply to a single disk.
No external storage devices detected.
No external storage modules (e.g. RA4000) were detected on the cluster. If external storage is actually used,
either the external storage device or the utilities to access it may have been installed incorrectly.
Disk Group "name" contains n dissociated plexes.
This is an informational message to inform the user that there are plexes in the named disk group which are
not associated with any disk volume. This may represent available storage space.
No Veritas mirrored volumes have been defined in name disk group.
There are no volumes in the specified disk group which contain mirrored plexes. This does not necessarily indicate
an error.
WARNING: Array containing / is controlled by only one node.
During the cross-referencing of Veritas and FCAL information, it was observed that a single node controls the
FCAL array containing the root filesystem. This introduces a single point of failure for the cluster.
Plex name contains both FCAL and non-FCAL disks.
The named plex consists of multiple subdisks, some under FCAL control and some not. This configuration is not
recommended and might represent an error.
Plex name has subdisks with different RAID configurations.
The named plex contains subdisks under FCAL control and those subdisks have different RAID (fault-tolerance)
definitions from the viewpoint of FCAL. This is not a recommended configuration and is probably an error.
Subdisk/disk-media information inconsistent.
The device reference in the subdisk record does not agree with the device reference in the disk media record
associated with the subdisk. This condition should never occur and probably indicates a problem with either the
Veritas configuration or with Veritas itself.
Volume: name mirrored on 3 or more plexes.
This is an informational message indicating that the specified volume contains more than two plexes. This does
not indicate an error; however, only two plexes are required for the mirroring of data.
WARNING - the volume containing / is not mirrored.
The volume containing the root filesystem is not mirrored. This message is not an error, but it does indicate
a single point of failure for the cluster.
Volume name uses both RAID and Veritas mirroring.
The volume named is mirrored by both the FCAL/RAID subsystem and by Veritas. This redundant mirroring may be
desirable in some extreme disaster recovery scenarios but it would not be used in most traditional configurations.
Volume name uses questionable mix of RAID and Veritas mirroring.
The mixture of RAID and Veritas mirroring in the named volume is highly questionable and probably indicates
a configuration error. The different plexes of the volume do not have the same RAID level.
name: Some subdisks in plex are controlled by different nodes.
For the named volume, the same cluster node does not control all of the subdisks that comprise a given plex.
While technically legal, this actually increases the probability of a given failure bringing down the plex, the
volume, and possibly the cluster as a whole.
WARNING - Volume name is cross-node mirrored in an FCAL environment.
This configuration is highly questionable. When FCAL is used to manage the disks, cross-node mirroring for
those same disks is not required nor recommended. The FCAL subsystem, if properly configured, will provide more
robust failover operations since remirroring of the volume would not be necessary in case of a node failure.
nscvu returns a non-0 value if the utility determines the cluster is in a non-compliant state or all information could not be obtained. This utility requires the nscsnmp SNMP support package and displays an error if the package is not properly installed.
casacfg(1M), cluster(1M), cron(1M)
Compaq ProLiant Clusters for SCO UnixWare 7 Reference Guide
Use Uninterruptible Power Supplies (UPSs) with Compaq ProLiant Clusters to minimize system downtime in the event of power loss. Cable the UPSs to enable the cluster to be shutdown cleanly before the UPS battery backup is exhausted. Proper cabling of UPSs is described in the ProLiant NonStop Clusters for SCO UnixWare 7 Reference Guide.
UPS-initiated shutdown can minimize data loss as well as improve cluster reboot speed when power returns. A monitoring process running within the cluster provides UPS-initiated shutdown. A simple configuration file controls this monitoring process.
Configure UPS-initiated shutdown by modifying the OS_SHUTDOWN_DELAY
, UPS_SERIAL_PORT
, and UPS_LOG_FILE
parameters
in the /opt/compaq/etc/nscupsd.cfg
configuration file.
OS_SHUTDOWN_DELAY
OS_SHUTDOWN_DELAY=15
indicates that a cluster-wide shutdown is initiated when the UPS has 15 minutes of battery backup power remaining. Measure the time required for a clean shutdown of
the cluster under peak operating conditions to ensure that the shutdown time is adequate.
UPS_LOG_FILE
var/spool/compaq/nscupsd.log
and should not be modified.
UPS_SERIAL_PORT
Colon-separated serial ports create a pair of UPSs in which both of the UPSs must signal that they are low on power before a cluster shuts down. Such a group of UPSs is referred to as a logical UPS. A logical UPS identifies UPSs that are grouped together to provide redundant power to a cluster. The drain of a single UPS within a logical UPS should not result in the loss of any key cluster resources.
Semicolon-separated serial ports identify a list of UPSs or a list of logical UPSs. A low-power indication by any of those UPSs results in a cluster-wide shutdown. This parameter is used when a cluster spans multiple power domains and the loss of any one domain results in a cluster-wide shutdown to protect the cluster.
NOTE: A serial connection to a UPS is used in determining shutdown if the node with the serial port is an active member in the cluster.
A two-node cluster or single-rack cluster with two UPSs, the cluster
should shut down when both UPSs are low on power.
To have both UPSs signal that they are low before shutting down the cluster, combine both UPSs into a single logical UPS. This results in a UPS_SERIAL_PORT
configuration of:
UPS_SERIAL_PORT="/dev/tty00.1:/dev/tty00.2"
/dev/tty00.1
is the device identifier for the node 1 serial port tied to UPS1 and
/dev/tty00.2
is the device identifier for the node 2 serial port tied to UPS2.
In a multiple-rack cluster that uses fully redundant
power in all key cluster components, the cluster should shut down only if both
UPSs in a single rack lose power. For both UPSs within a rack to signal
that they are low before shutting down the cluster, both UPSs within a rack are
joined into a single logical UPS. This results in an UPS_SERIAL_PORT
configuration of:
example: UPS_SERIAL_PORT="/dev/tty00.1:/dev/tty00.2;/dev/tty01.1:/dev/tty01.2;"
After UPS-initiated shutdown has been configured as described in the preceding sections, restart the UPS daemon by entering the following two commands:
/etc/init.d/nscups stop
/etc/init.d/nscups start
If multiple UPSs are configured and one UPS loses power, the warning message will incorrectly state the system will shut down when UPS1 and UPS1 lose power instead of stating the system will shut down when UPS1 and UPS2 lose power. The system will operate correctly, the message is in error.
g sections, restart the UPS daemon by entering the following two commands:
/etc/init.d/nscups stop
/etc/init.d/nscups start