Errors

Nemeth Lorant <[email protected]> Tue, 7 May 2002 15:03:43 +0200 (CEST)
Newsgroups gmane.linux.failsafe
Message-ID <[email protected]>
Hi!

I'm trying to bring up a two node cluster, but I have several errors. The
machines are runnig redhat 7.2. I build the RPMs with the recomended
patches (glibc 2.2.4,STONITH). After cdbreinit, I defined two nodes, a
cluster, 2 policies, 2 IP_addresses, 2 resource groups. The corresponding
scripts may be seen in the attachments. After that is started ha sevices 
from the CLI (start ha_services), but the 2 IP addresses didn't get 
available. I pasted some messeges that semms to show some errors. Any 
ideas what the problem could be. I would be glad for any help, because 
next week I'll have to give a talk about failsafe.


			Thx:

					Lorant Nemeth (Loci)


--- cad_log --

Tue May 7 12:02:24.362 <cad 1438:7176> cfs_fs_connect: fs_cam_register
failed with error FailSafe is not ready to accept admin requests.

Tue May 7 12:02:34.362 <cad 1438:7176> cfs_fs_connect: fs_cam_register 
failed with error FailSafe is not ready to accept admin requests.

Tue May 7 12:02:44.362 <cad 1438:7176> cfs_fs_connect: fs_cam_register 
failed with error FailSafe is not ready to accept admin requests.

--- failsafe_labpc1 ---

Tue May  7 11:30:05.741 <D0 ha_fsd config 2132:0 ci_cdbconfig_nodes.c:217> 
Configuration read successful.
Tue May 7 11:30:05.746 <D0 ha_fsd fsd 2132:0 fs_main.c:227> Initial sleep 
time 60 seconds

Tue May 7 11:31:05.754 <W ha_fsd ipc 2132:0 ipc_clnt.c:388>
CI_IPCERR_NOSERVER, ipcclnt_connect(): server pid 1850 in connection file
/var/run/failsafe/comm/crsd-ipc_labpc1 is not valid. 
Tue May 7 11:31:07.761 <W ha_fsd fsd 2132:0 fs_crs.c:71> 
CI_IPCERR_NOSERVER, Waiting for crsd to accept connections


--- ifd_labpc1 ---

                * * * L o g g i n g R e s t a r t e d * * *
Tue May 7 11:30:09.458 <W ha_ifd ifd 2127:0 ifd_net.c:679> CI_FAILURE, lo 
is neither broadcast, point-to-point, nor loopback


--- srmd_labpc1 --- 
Tue May 7 14:06:52.445 <I0 ha_srmd srm 2164:2 sc_resource.c:2492> Resource 
name 152.66.242.65 type IP_address
Tue May 7 14:06:52.446 <I0 ha_srmd srm 2164:2 sc_reply.c:133> Resource 
information request reply done
Tue May 7 14:06:52.447 <I0 ha_srmd srm 2164:2 sc_request.c:130> Client 
unregister request
Tue May 7 14:06:52.448 <I0 ha_srmd srm 2164:2 sc_reply.c:133> Unregister 
request reply done

----

cmgr> show resources in resource_group ResourceGroup1

Resources: 
	152.66.242.65	(type: IP_address)

cmgr> show status of resource 152.66.242.65 of resource_type IP_address

Status: no such resource

----

cmgr> admin online resource_group ResourceGroup1
FailSafe daemon (ha_fsd) is not running on this local node or it is not 
ready to accept admin commands.
Resource Group (ResourceGroup1) is online-ready.

Failed to admin:
	online

admin command failed
cmgr> show status of resource_group ResourceGroup1

State: Online Ready
Error: No error
Check resource group ResourceGroup1 status in an active node if HA 
services are active in cluster
cmgr-create-resource-IP_addresses (text/plain, 2.2 KB)
#
# The following cluster_mgr script allows users to define a resource of 
# resource type IP_address for a cluster in the CDB. This script has to run as 
# "/usr/cluster/bin/cluster_mgr -f".  Replace all the values that are
# within "<<" and ">>" with values appropriate for the resource and 
# run the script.
#

#
# If multiple resources of resource type IP_address have to be defined, repeat
# the following IP_address definition template.
#
# Name of the IP_address resource.  The name of the resource has to be IP 
# address in the internet "." notation. This IP address is used by 
# clients to access highly-available resources.
# Example value: 192.26.50.140
#
define resource 152.66.242.65 of resource_type IP_address in cluster HA

#
# The network mask for the IP address. The network mask value is used
# to configure the IP address on the network interface.
# Example value: 0xffffff00
	set NetworkMask to 0xffffffe0
# 
# The ordered list of interfaces that can be used to configure the IP address.
# The list of interface names are separated by comma. 
# Example value: ef0, ef1
	set interfaces to eth0
#
# The broadcast address for the IP address.
# Example value: 192.26.50.255
	set BroadcastAddress to 152.66.242.95

#
# IP_address resource definition for the cluster complete
#
done

#
# Name of the IP_address resource.  The name of the resource has to be IP 
# address in the internet "." notation. This IP address is used by 
# clients to access highly-available resources.
# Example value: 192.26.50.140
#
define resource 152.66.242.66 of resource_type IP_address in cluster HA

#
# The network mask for the IP address. The network mask value is used
# to configure the IP address on the network interface.
# Example value: 0xffffff00
	set NetworkMask to 0xffffffe0
# 
# The ordered list of interfaces that can be used to configure the IP address.
# The list of interface names are separated by comma. 
# Example value: ef0, ef1
	set interfaces to eth0
#
# The broadcast address for the IP address.
# Example value: 192.26.50.255
	set BroadcastAddress to 152.66.242.95

#
# IP_address resource definition for the cluster complete
#
done

#
# Script complete. This should be last line of the script
#
quit
cmgr-create-resource_group (text/plain, 2.3 KB)
#
# The following cluster_mgr script allows users to define a resource group in 
# a cluster in the CDB. This script has to run as 
# "/usr/lib/failsafe/bin/cluster_mgr -f".  Replace all the values that are 
# within "<<" and ">>" with values appropriate for the failover policy and 
# run the script.
#

#
# Name of the resource group. Name of the resource group must be unique in 
# the cluster.
#
define resource_group ResourceGroup1 in cluster HA
#
# Failover policy for the resource group. This field is mandatory. 
# Failover policy should be already defined in the CDB.
#
	set failover_policy to policy1
#
# List of resources in the resource group.
# Repeat the following line for each resource to be added to the resource group.
#
	add resource 152.66.242.65 of resource_type IP_address
#	add resource <<resource name>> of resource_type <<resource type name>>
#	add resource <<resource name>> of resource_type <<resource type name>>
#	add resource <<resource name>> of resource_type <<resource type name>>
# Add more resources to the resource group here.
#

#
# Resource group definition complete
#
done

# The following cluster_mgr script allows users to define a resource group in 
# a cluster in the CDB. This script has to run as 
# "/usr/lib/failsafe/bin/cluster_mgr -f".  Replace all the values that are 
# within "<<" and ">>" with values appropriate for the failover policy and 
# run the script.
#

#
# Name of the resource group. Name of the resource group must be unique in 
# the cluster.
#
define resource_group ResourceGroup2 in cluster HA
#
# Failover policy for the resource group. This field is mandatory. 
# Failover policy should be already defined in the CDB.
#
	set failover_policy to policy2
#
# List of resources in the resource group.
# Repeat the following line for each resource to be added to the resource group.
#
	add resource 152.66.242.66 of resource_type IP_address
#	add resource <<resource name>> of resource_type <<resource type name>>
#	add resource <<resource name>> of resource_type <<resource type name>>
#	add resource <<resource name>> of resource_type <<resource type name>>
# Add more resources to the resource group here.
#

#
# Resource group definition complete
#
done

#
# Script complete. This should be last line of the script
#
quit
cmgr-create-failover_policy2 (text/plain, 1.1 KB)
#
# The following cluster_mgr script allows users to define a failover policy in 
# the CDB. This script has to run as "/usr/lib/failsafe/bin/cluster_mgr -f".
# Replace all the values that are within "<<" and ">>" with values appropriate
# for the cluster and run the script.
#

#
# Name of the failover policy.  
#
define failover_policy policy2
#
# Failover policy attribute. This field is mandatory. 
# Possible values: Auto_Failback, Controlled_Failback, Auto_Recovery,
#		InPlace_Recovery
#
	set attribute to Auto_Failback
# 
# Failover policy script. The failover policy scripts have to be present in
# /usr/lib/failsafe/policies directory. This field is mandatory.
# Example value: ordered (file name not the full path name).
	set script to ordered
#
# Failover policy domain. Ordered list of nodes in the cluster separated by
# spaces. This field is mandatory.
#
	set domain to  labpc2.inf.mit.bme.hu labpc1.inf.mit.bme.hu
#
# Failover policy definition complete
#
done

#
# Script complete. This should be last line of the script
#
quit
cmgr-create-failover_policy1 (text/plain, 1.1 KB)
#
# The following cluster_mgr script allows users to define a failover policy in 
# the CDB. This script has to run as "/usr/lib/failsafe/bin/cluster_mgr -f".
# Replace all the values that are within "<<" and ">>" with values appropriate
# for the cluster and run the script.
#

#
# Name of the failover policy.  
#
define failover_policy policy1
#
# Failover policy attribute. This field is mandatory. 
# Possible values: Auto_Failback, Controlled_Failback, Auto_Recovery,
#		InPlace_Recovery
#
	set attribute to Auto_Failback
# 
# Failover policy script. The failover policy scripts have to be present in
# /usr/lib/failsafe/policies directory. This field is mandatory.
# Example value: ordered (file name not the full path name).
	set script to ordered
#
# Failover policy domain. Ordered list of nodes in the cluster separated by
# spaces. This field is mandatory.
#
	set domain to  labpc1.inf.mit.bme.hu labpc2.inf.mit.bme.hu
#
# Failover policy definition complete
#
done

#
# Script complete. This should be last line of the script
#
quit
cmgr-create-cluster (text/plain, 1.6 KB)
#
# The following cluster_mgr script allows users to define a cluster in the
# CDB. This script has to run as "/usr/cluster/bin/cluster_mgr -f".
# Replace all the values that are within "<<" and ">>" with values appropriate
# for the cluster and run the script.
#

#
# Name of the cluster.  
#
define cluster HA
#
#	Creating a failsafe cluster
# Following line is mandatory for FailSafe clusters
	set is_failsafe to true
#
# Notification command for the cluster. This is optional. If this field is
# not specified,  /usr/bin/mail command is used for notification. Notification
# is sent when there is change in status of cluster, node and resource group.
#
#	set notify_cmd to <<command path>>
# 
# Notification address for the cluster. This field value is passed as
# argument to the notification command. This field is optional. If notification
# address is not provided, notification command should not require an
# notification address argument.  If both notification address and
# notification command are not specified, notification are sent for this
# cluster.
# Example value: [email protected]
	set notify_addr to fsafe_admin
#
# List of nodes added to the cluster.
# Repeat the following line for each node to be added to the cluster.
# Node should be already defined in the CDB and logical name of the
# node has to be specified.
	add node labpc1.inf.mit.bme.hu
	add node labpc2.inf.mit.bme.hu
#
# Add more nodes to the cluster here.
#

#
# Cluster definition complete
#
done

# Set default cluster

set cluster HA
#
# Script complete. This should be last line of the script
#
quit
cmgr-create-node2 (text/plain, 5 KB)
#
# The following cluster_mgr script allows users to define a node in the
# CDB. This script has to run as "/usr/cluster/bin/cluster_mgr -f".
# Replace all the values that are within "<<" and ">>" with values appropriate
# for the node and run the script.
#

#
# logical name of the node. It is recommended that logical name of the node
# be output of hostname(1) command.
#
define node labpc2.inf.mit.bme.hu
#
# Hostname of the node. This is optional. If this field is not specified,
# logical name of the node is assumed to be hostname. This value has to be
# the output of hostname(1) command.
#
#	set hostname to labpc2.inf.mit.bme.hu
#
#	Creating a failsafe node
# Following line is mandatory for node that is part of FailSafe clusters
	set is_failsafe to true
# 
# Node identifier. Node identifier is a 16 bit integer that uniquely
# identifies the node. This field is optional. If value is not provided,
# cluster software generates node identifier.
# Example value: 1
	set nodeid to 2
#
# Description of the system controller of this node.
# System controller can be "chalL" or "msc" or "mmsc". If the node is a
# Challenge DM/L/XL, then system controller type is "chalL". If the node 
# is Origin 200 or deskside Origin 2000, then the system controller type 
# is "msc". If the node is rackmount Origin 2000, the system controller
# type is "mmsc".
# Possible values: msc, mmsc, chalL
#
	set sysctrl_type to stonith
#
# You can enable or disable system controller definition. Users are 
# expected to enable system controller definition after verify the
# serial reset cables connected to this node.
# Possible values: enabled, disabled
#
	set sysctrl_status to enabled
# 
# The system controller password for doing privileged system controller
# commands.
# This field is optional.
#
#	set sysctrl_password to <<system_controller_password>>
#
# System controller owner. The node name of the machine that is 
# connected using serial cables to system controller of this node.
# System controller node also has to be defined in the CDB.
#
	set sysctrl_owner to labpc1.inf.mit.bme.hu
#
# System controller device. The absolute device path name of the tty to which
# the serial cable is connected in this node.
# Example value: /dev/ttyd2
#
	set sysctrl_device to /dev/ttyS0
#
# Currently, the system controller owner can be connected to the system
# controller on this node using "tty" device.  
# Possible value: tty
#
	set sysctrl_owner_type to tty
#
# List of control networks. There can be multiple control networks specified
# for a node. HA cluster software uses these control networks for 
# communication between nodes.  At least two control networks should 
# be specified for heartbeat messages and one control network for
# FailSafe control messages.
# For each control network for the node, please add one more control network
# section.
#
# Name of control network IP address. This IP address must be configured on
# the network interface in /etc/config/netif.options file in the node.
# It is recommended that the IP address in internet dot notation is provided.
# Example value: 192.26.50.3
#
	add nic 152.66.242.72
#
# Flag to indicate if the control network can be used for sending 
# heartbeat messages.
# Possible values: true, false
#
	    set heartbeat to true
#
# Flag to indicate if the control network can be used for sending 
# FailSafe control messages.
# Possible values: true, false
#
	    set ctrl_msgs to true
#
# Priority of the control network. Higher the priority value, lower the
# priority of the control network.
# Example value: 1
#
	    set priority to 1
#
# Control network information complete
#
	done
#
# List of control networks. There can be multiple control networks specified
# for a node. HA cluster software uses these control networks for 
# communication between nodes.  At least two control networks should 
# be specified for heartbeat messages and one control network for
# FailSafe control messages.
# For each control network for the node, please add one more control network
# section.
#
# Name of control network IP address. This IP address must be configured on
# the network interface in /etc/config/netif.options file in the node.
# It is recommended that the IP address in internet dot notation is provided.
# Example value: 192.26.50.3
#
	add nic 192.168.0.2
#
# Flag to indicate if the control network can be used for sending 
# heartbeat messages.
# Possible values: true, false
#
	    set heartbeat to true
#
# Flag to indicate if the control network can be used for sending 
# FailSafe control messages.
# Possible values: true, false
#
	    set ctrl_msgs to true
#
# Priority of the control network. Higher the priority value, lower the
# priority of the control network.
# Example value: 1
#
	    set priority to 0
#
# Control network information complete
#
	done
#
# Add more control networks information here.
#

#
# Node definition complete
#
done

#
# Script complete. This should be last line of the script
#
quit
cmgr-create-node1 (text/plain, 5 KB)
#
# The following cluster_mgr script allows users to define a node in the
# CDB. This script has to run as "/usr/cluster/bin/cluster_mgr -f".
# Replace all the values that are within "<<" and ">>" with values appropriate
# for the node and run the script.
#

#
# logical name of the node. It is recommended that logical name of the node
# be output of hostname(1) command.
#
	define node labpc1.inf.mit.bme.hu
#
# Hostname of the node. This is optional. If this field is not specified,
# logical name of the node is assumed to be hostname. This value has to be
# the output of hostname(1) command.
#
#set hostname to labpc1.inf.mit.bme.hu
#
#	Creating a failsafe node
# Following line is mandatory for node that is part of FailSafe clusters
	set is_failsafe to true
# 
# Node identifier. Node identifier is a 16 bit integer that uniquely
# identifies the node. This field is optional. If value is not provided,
# cluster software generates node identifier.
# Example value: 1
	set nodeid to 1
#
# Description of the system controller of this node.
# System controller can be "chalL" or "msc" or "mmsc". If the node is a
# Challenge DM/L/XL, then system controller type is "chalL". If the node 
# is Origin 200 or deskside Origin 2000, then the system controller type 
# is "msc". If the node is rackmount Origin 2000, the system controller
# type is "mmsc".
# Possible values: msc, mmsc, chalL
#
	set sysctrl_type to stonith
#
# You can enable or disable system controller definition. Users are 
# expected to enable system controller definition after verify the
# serial reset cables connected to this node.
# Possible values: enabled, disabled
#
	set sysctrl_status to enabled
# 
# The system controller password for doing privileged system controller
# commands.
# This field is optional.
#
#	set sysctrl_password to none
#
# System controller owner. The node name of the machine that is 
# connected using serial cables to system controller of this node.
# System controller node also has to be defined in the CDB.
#
	set sysctrl_owner to labpc2.inf.mit.bme.hu
#
# System controller device. The absolute device path name of the tty to which
# the serial cable is connected in this node.
# Example value: /dev/ttyd2
#
	set sysctrl_device to /dev/ttyS0
#
# Currently, the system controller owner can be connected to the system
# controller on this node using "tty" device.  
# Possible value: tty
#
	set sysctrl_owner_type to tty
#
# List of control networks. There can be multiple control networks specified
# for a node. HA cluster software uses these control networks for 
# communication between nodes.  At least two control networks should 
# be specified for heartbeat messages and one control network for
# FailSafe control messages.
# For each control network for the node, please add one more control network
# section.
#
# Name of control network IP address. This IP address must be configured on
# the network interface in /etc/config/netif.options file in the node.
# It is recommended that the IP address in internet dot notation is provided.
# Example value: 192.26.50.3
#
	add nic 152.66.242.71
#
# Flag to indicate if the control network can be used for sending 
# heartbeat messages.
# Possible values: true, false
#
	    set heartbeat to true
#
# Flag to indicate if the control network can be used for sending 
# FailSafe control messages.
# Possible values: true, false
#
	    set ctrl_msgs to true
#
# Priority of the control network. Higher the priority value, lower the
# priority of the control network.
# Example value: 1
#
	    set priority to 1
#
# Control network information complete
#
	done
#
# List of control networks. There can be multiple control networks specified
# for a node. HA cluster software uses these control networks for 
# communication between nodes.  At least two control networks should 
# be specified for heartbeat messages and one control network for
# FailSafe control messages.
# For each control network for the node, please add one more control network
# section.
#
# Name of control network IP address. This IP address must be configured on
# the network interface in /etc/config/netif.options file in the node.
# It is recommended that the IP address in internet dot notation is provided.
# Example value: 192.26.50.3
#
	add nic 192.168.0.1
#
# Flag to indicate if the control network can be used for sending 
# heartbeat messages.
# Possible values: true, false
#
	    set heartbeat to true
#
# Flag to indicate if the control network can be used for sending 
# FailSafe control messages.
# Possible values: true, false
#
	    set ctrl_msgs to true
#
# Priority of the control network. Higher the priority value, lower the
# priority of the control network.
# Example value: 1
#
	    set priority to 0
#
# Control network information complete
#
	done
#
# Add more control networks information here.
#

#
# Node definition complete
#
done

#
# Script complete. This should be last line of the script
#
quit