 |
Index for
Section 8 |
|
 |
Alphabetical
listing for C |
|
 |
Bottom of
page |
|
cnxshow(8)
NAME
cnxshow - Verifies the correct operation of the cluster communications
subsystem and displays the connection manager's cluster view (provided on
Production Server configurations only)
SYNOPSIS
/usr/sbin/cnxshow [-i interval] [-c] [-d] [-m] [-n name] [-v]
FLAGS
-i interval
Specifies the frequency (in seconds) with which the cnxshow
utility updates and displays the cluster view. By default, the
utility does not update the display. To stop the display, press
Ctrl/C.
-c Uses the curses library to display information.
-d Activates verbose mode when the -v flag is also specified. In
verbose mode, the cnxshow utility provides information about the
progress and outcome of the connection manager (CNX) Installation
Verification Procedure (IVP) tests.
-m Displays the status of the connection manager's monitor daemon
(cnxmond).
-n name Displays the specified member system's cluster view. By default,
the cluster view of the current system is displayed.
-v Runs the CNX IVP tests. These tests verify that the connection
manager director (cnxmgrd) exists and is active, confirm cluster
membership, note the integrity of the distributed lock manager's
(DLM's) communications path to each cluster member, and verify
that the clocks on each member system are properly synchronized.
DESCRIPTION
The cnxshow utility reports which systems are members of the cluster,
validates that the system clocks of cluster member systems are
synchronized, and determines whether the cluster communications subsystem
is operating properly on each system. It is called by the cluster IVP
utility, clu_ivp(8); you can also invoke it as a separate utility.
The cnxshow utility also displays the connection manager's cluster view,
providing the following information:
Director: Memory Channel IP name (as defined by the CLUSTER_NET
configuration variable in the /etc/rc.config file) of the member
system currently running the cluster director daemon (cnxmgrd).
Suspended:
State of the cluster director daemon. The value No indicates a
normal state. When the value of Suspended is Yes, the
distributed lock manager (DLM) has been temporarily disabled
because the member system is not in contact with the cluster
director (cnxmgrd). In this case, it is impossible for the
cnxshow utility to determine whether the system is a cluster
member or not, so it puts a question mark (?) in the Member
column in the display.
Node monitor using tie-breaking disk:
If the cluster contains two members and is using its Memory
Channel interconnect in virtual hub mode, this item lists the
names of the raw devices the active cluster monitor daemon
(cnxmond) is using as tie-breaker disks. In such a cluster, you
invoke the cnxset utility on both members to configure one, two,
or three tie-breaker disks. The names of the tie-breaker disks
are stored locally in the CNX_DISK configuration variable in each
member's /etc/rc.config file.
If the tie-breaker disks used by the cluster monitor differ from
those specified in the local member's /etc/rc.config file, the
utility displays these additional messages:
The node monitor and the local node specify different disks.
The local node specifies disk : <disk>
In this event, use the cnxset utility to ensure that both cluster
members define the same tie-breaker disks.
If a two-member cluster is running in virtual hub mode but no
tie-breaker disk has been specified, the cnxshow utility displays
the following message:
Node monitor using tie-breaking disk: Required but not defined
If the cluster is configured to use a Memory Channel hub, the
utility displays the following:
Node monitor using tie-breaking disk: Not required
Hostname Host name of each member system (as defined by the HOSTNAME
configuration variable in each system's /etc/rc.config file).
Cluster I/F
Memory Channel IP name of each member system (as defined by the
CLUSTER_NET configuration variable in each system's
/etc/rc.config file).
CS_ID Cluster system ID of each member system. The CS_ID identifies
the member system in the cluster configuration information
maintained by the connection manager.
Incarnation
Incarnation number of each member system. A member system's
incarnation number changes when it is rebooted, and thus
indicates the current instance of a running system and a change
in its software context.
Comm Okay State of the distributed lock manager's (DLM's) communications
path. The value Yes indicates a normal state.
Member Whether the listed system is currently a cluster member. The
value in this column is a question mark (?) when the value in the
Suspended column is Yes.
When you specify the -m flag, the cnxshow utility displays the status of
the active cluster monitor daemon. This display includes the following
information:
Virtual hub
Whether the cluster uses its Memory Channel interconnect in
virtual hub mode. A two-member cluster typically uses virtual
hub mode.
Tie-breaker disk status
Listing of the one, two, or three disks that have been configured
as tie-breaker disks in a cluster using the Memory Channel
interconnect in virtual hub mode. The display indicates how many
disks are needed to maintain the cluster when only one member
system remains. The open: and error: fields provide no useful
information. They are reserved for Compaq internal use.
Node Memory Channel IP name of each member system (as defined by the
CLUSTER_NET configuration variable in each system's
/etc/rc.config file).
Incarnation
Incarnation number of each member system. A member system's
incarnation number changes when it is rebooted, and thus
indicates the current instance of a running system and a change
in its software context.
State State of each member system. In normal operation, a member's
state is either mbr_member or mbr_not_member.
Acked Reserved for Compaq internal use.
Ping_Exptd
Reserved for Compaq internal use.
If the system on which you execute the cnxshow utility is disconnected from
the Memory Channel interconnect, the utility displays the following
message:
Error getting cluster view - Please check Memory Channel
EXAMPLES
In the following example, the cnxshow utility is run on system mcclu14 in a
cluster running in virtual hub mode:
% /usr/sbin/cnxshow
Cluster View from mcclu14
Director: mcclu13, Suspended: No
Node monitor using tie-breaking disk: /dev/rrz10c
Hostname Cluster I/F CS_ID Incarnation Comm Okay Member
-----------------------------------------------------------------------------
clu14 mcclu14 0001,0002 000000000006a090 Yes Yes
clu13 mcclu13 0001,0001 0000000000062e30 Yes Yes
In the following example, the cnxshow utility is run with the -m flag on
system mcclu8 in a cluster running in virtual hub mode:
% /usr/sbin/cnxshow -m
Node Monitor Status
Virtual hub: Yes
Tie-breaker disk status (need 2 of 3) when one node present
/dev/rrz10c open: Yes error: No
/dev/rrz11c open: No error: Yes
/dev/rrz12c open: Yes error: No
Node Incarnation State Acked Ping_Exptd
-----------------------------------------------------------------------------
mcclu14 000000000006a090 mbr_member Yes Yes
mcclu13 0000000000062e30 mbr_member Yes Yes
The following command runs the CNX IVP tests on a cluster:
% cnxshow -d -v
Successfully verified that a CNX director exists.
Successfully verified that the CNX director is not suspended.
Successfully verified CNX comm state of pikerm.
Successfully verified that pikerm is a CNX member.
Successfully verified CNX comm state of sparkyrm.
Successfully verified that sparkyrm is a CNX member.
Successfully verified that clocks on sparkyrm and pikerm are within 15 seconds.
CNX ivp tests exiting successfully with no errors found.
FILE
/usr/sbin/cnxshow
Specifies the command path.
RELATED INFORMATION
cnxagentd(8), cnxmgrd(8), cnxmond(8), cnxpingd(8), cnxset(8)
|
Index for
Section 8 |
|
|
Alphabetical
listing for C |
|
 |
Top of
page |
|