03-09-2017 02:27 AM
hi,
i noticed that after i start a group and a resource within this group is not brought online,after i bring it online with OS commands,the resource remains online at OS level after i stop vcs.
I think it is good to verify the resources at OS level after the vcs is down.right?
VCS could not bring the resource down because it was not brought online by vcs.right?
tnx
Solved! Go to Solution.
03-13-2017 04:33 PM
This configuration is incomplete - to bring a zpool up you need to configure a Zpool agent and add the resulting Zpool resource as a child dependency to the mount point with ZFS.
Have a look at the Cluster Server 7.2 Bundled Agents Reference Guide - Solaris
03-09-2017 02:12 PM
Hi,
This does not sound right. If a non-persistent resource is not brought up upon service group startup then there is an issue with configuration of that resource. So it is likely not properly started, stopped and monitored by VCS.
Otherwise, depending how you stop VCS (using hastop or hastop -all -force ) - it will or will not shut down service groups, respectively.
03-09-2017 11:02 PM
hi,
i have:
root@test89:/etc/VRTSvcs/conf/config# more main.cf
include "OracleASMTypes.cf"
include "types.cf"
include "Db2udbTypes.cf"
include "OracleTypes.cf"
include "SybaseTypes.cf"
cluster clust (
ClusterAddress = "192.168.114.35"
SecureClus = 1
GuestGroups = { test5 }
)
system test89 (
)
group nic (
SystemList = { test89 = 0 }
AutoStartList = { test89 }
)
NIC test5n (
Device = net0
NetworkType = ether
)
Phantom grp1phantom (
)
// resource dependency tree
//
// group nic
// {
// Phantom grp1phantom
// NIC test5n
// }
group test5 (
SystemList = { test89 = 0 }
AutoStartList = { test89 }
)
Application apache (
StartProgram = "/usr/apache2/2.2/bin/apachectl start"
StopProgram = "/lib/svc/method/http-apache22 stop"
CleanProgram = "/root/test5"
PidFiles = { "/system/volatile/apache2/2.2/httpd.pid" }
)
IP IP_RES (
Device = net0
Address = "192.168.114.88"
NetMask = "255.255.255.0"
IfconfigTwice = 1
)
Mount mnt (
MountPoint = "/n"
BlockDevice = "t5/n"
FSType = zfs
FsckOpt = "[%-n]"
)
Proxy nicproxy (
TargetResName = test5n
)
IP_RES requires nicproxy
apache requires IP_RES
root@test89:/etc/VRTSvcs/conf/config# hagrp -state
#Group Attribute System Value
nic State test89 |ONLINE|
test5 State test89 |PARTIAL|STARTING|
root@test89:/etc/VRTSvcs/conf/config# hares -state
#Resource Attribute System Value
IP_RES State test89 ONLINE
apache State test89 ONLINE
grp1phantom State test89 ONLINE
mnt State test89 FAULTED
nicproxy State test89 ONLINE
test5n State test89 ONLINE
mnt is not online.
/var/VRTSvcs/log/engine_A.log
2017/03/11 08:53:45 VCS INFO V-16-1-10304 Resource mnt (Owner: Unspecified, Group: test5) is offline on test89 (First probe)
2017/03/11 08:54:25 VCS NOTICE V-16-1-10301 Initiating Online of Resource mnt (Owner: Unspecified, Group: test5) on System test89
2017/03/11 08:56:26 VCS ERROR V-16-2-13066 (test89) Agent is calling clean for resource(mnt) because the resource is not up even after online completed.
2017/03/11 08:56:28 VCS INFO V-16-2-13068 (test89) Resource(mnt) - clean completed successfully.
2017/03/11 08:56:28 VCS INFO V-16-2-13071 (test89) Resource(mnt): reached OnlineRetryLimit(0).
2017/03/11 08:56:28 VCS ERROR V-16-1-54031 Resource mnt (Owner: Unspecified, Group: test5) is FAULTED on sys test89
tnx
03-10-2017 01:34 AM
hi,
if zpool test88 is imported then resource mnt is online.
i thought that vcs will import the zpool and mount to /n but i see that test88 must be imported and vcs mount to /n.
is this the expected behavior?
tnx
03-13-2017 04:33 PM
This configuration is incomplete - to bring a zpool up you need to configure a Zpool agent and add the resulting Zpool resource as a child dependency to the mount point with ZFS.
Have a look at the Cluster Server 7.2 Bundled Agents Reference Guide - Solaris