topic Re: cmrunpkg : Script failed with no restart in Operating System - Linux

cmrunpkg : Script failed with no restart

NiCK_76 — Wed, 07 Dec 2005 11:03:58 GMT

Hi expert,

I never,ever install service guard for linux before. Now, I encounted a problem when I perform cmrunpkg for apache toolkit. The error messages as following:
[root@sglx02 http_pkg1]# cmrunpkg -n sglx02 http_pkg1
cmrunpkg : Script failed with no restart: http_pkg1 should not be restarted.
Check the syslog on node sglx02 and pkg log files for more detailed information.

My system information as following:
node 1 : HP DL580 G3 with 2 HBAs
node 2 : HP DL380 G3 with 2 HBAs
storage: HP MSA 1000 with 1 san 8 port switch
os : RHEL AS3 U6
service guard for linux 11.16
apache toolkit A.02.00

Thank you for any advance

Re: cmrunpkg : Script failed with no restart

John Bigg — Wed, 07 Dec 2005 11:24:11 GMT

This error indicates that the package tried to start but encountered a fatal error. You need to look at the package log file to determine what that error was. The log file will probably be located somewhere under $SGCONF (/usr/local/cmcluster/conf for RHEL) although it can be any directory which was specified when the package was configured.

You can use cmviewconf to determine the exact location of the log files since they will be located in the same directory as the run script file which is listed by this command.

Re: cmrunpkg : Script failed with no restart

NiCK_76 — Wed, 07 Dec 2005 11:34:15 GMT

Hi John,

Thank you for you replay quickly.The following is cmviewconf result.I can't find any log messages.
[root@sglx02 http_pkg1]# cmviewconf

Cluster information:

cluster name: cluster1
version: 0
flags: 50
heartbeat interval: 1.00 (seconds)
node timeout: 2.00 (seconds)
heartbeat connection timeout: 4.00 (seconds)
auto start timeout: 600.00 (seconds)
network polling interval: 2.00 (seconds)
qs host: (not configured)

Cluster Node information:

Node ID 1:
Node name: sglx01
cluster lock lun name: /dev/sda1

Network ID 1:
ppa: 0
old_ppa: 0
mac addr: 0x0014c2633d41
hardware path: eth0
network interface name: eth0

IPv4 Information:
subnet: 192.168.8.0
subnet mask: 255.255.248.0
ip address: 192.168.15.115

IPv6 Information:

flags: 0 (Non-Heartbeat Network)

Network ID 2:
ppa: 0
old_ppa: 0
mac addr: 0x0014c2633d40
hardware path: eth1
network interface name: eth1

IPv4 Information:
subnet: 10.1.1.0
subnet mask: 255.255.255.0
ip address: 10.1.1.1

IPv6 Information:

flags: 1 (Heartbeat Network)

Node ID 2:
Node name: sglx02
cluster lock lun name: /dev/sda1

Network ID 1:
ppa: 0
old_ppa: 0
mac addr: 0x001321ccf14d
hardware path: eth0
network interface name: eth0

IPv4 Information:
subnet: 192.168.8.0
subnet mask: 255.255.248.0
ip address: 192.168.15.116

IPv6 Information:

flags: 0 (Non-Heartbeat Network)

Network ID 2:
ppa: 0
old_ppa: 0
mac addr: 0x001321ccf14c
hardware path: eth1
network interface name: eth1

IPv4 Information:
subnet: 10.1.1.0
subnet mask: 255.255.255.0
ip address: 10.1.1.2

IPv6 Information:

flags: 1 (Heartbeat Network)

Cluster Access Policy Information: (Not Defined)

Package information:

maximum configured packages: 30

package ID 29185:
package name: http_pkg1
package global flags: 1
(Package Switch Enabled)
(Configured Node Failover)
(Manual Failback)
package run script: /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl
package run timeout: (No Timeout)
package halt script: /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl
package halt timeout: (No Timeout)
package primary node: sglx01
package alternate node: sglx02

package services:
service ID: 1
service name: httpd.monitor
service halt timeout: 0 (seconds)
service fail fast: Disabled

package access policies: (Not Defined)

Re: cmrunpkg : Script failed with no restart

Luk Vandenbussche — Wed, 07 Dec 2005 12:33:58 GMT

Hi nick

Try this

cmmodpkg -e -n sglx02 http_pkg1
cmrunpkg -n sglx02 http_pkg1

Re: cmrunpkg : Script failed with no restart

melvyn burnard — Wed, 07 Dec 2005 14:21:04 GMT

Does this failure occur on just one node? or both nodes?

What does the /var/log/messages file contain regarding this failure?
What does it say in the package logs on the node it failed on? look in:
/usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl.log

Re: cmrunpkg : Script failed with no restart

NiCK_76 — Wed, 07 Dec 2005 21:37:02 GMT

Hi Luk,

It's still not working.
[root@sglx02 root]# cmmodpkg -e -n sglx02 http_pkg1
cmmodpkg : Warning: Node sglx02 is already able to run package http_pkg1.
cmmodpkg : Completed successfully on all packages specified.
[root@sglx02 root]# cmrunpkg -n sglx02 http_pkg1
cmrunpkg : Script failed with no restart: http_pkg1 should not be restarted.
Check the syslog on node sglx02 and pkg log files for more detailed information.

Re: cmrunpkg : Script failed with no restart

NiCK_76 — Wed, 07 Dec 2005 21:46:47 GMT

Hi melvyn burnard

The failure occur on both nodes.

The log file 'messages' contain following messages
Dec 8 10:41:38 sglx02 CM-CMD[22801]: cmrunpkg -n sglx02 http_pkg1
Dec 8 10:41:38 sglx02 CM-CMD[22801]: Request from root on node sglx02 to start package
Dec 7 18:41:38 sglx02 cmcld[6007]: Request from node sglx02 to start package http_pkg1 on node sglx02.
Dec 7 18:41:38 sglx02 cmcld[6007]: Executing '/usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl start' for package http_pkg1, as service PKG*29185.
Dec 8 10:41:38 sglx02 modprobe: modprobe: Can't locate module block-major-43
Dec 8 10:41:39 sglx02 last message repeated 31 times
Dec 7 18:41:39 sglx02 cmcld[6007]: Service PKG*29185 terminated due to an exit(1).
Dec 7 18:41:39 sglx02 cmcld[6007]: Package http_pkg1 run script exited with NO_RESTART.
Dec 7 18:41:39 sglx02 cmcld[6007]: Examine the file /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl.log for more details.
Dec 7 18:41:39 sglx02 cmcld[6007]: Switching disabled on package http_pkg1.
Dec 7 18:41:39 sglx02 cmcld[6007]: Unable to start package http_pkg1. Node sglx02 is not able to run it.

The log file 'http_pkg.cntl.log' contain nothing.

Thankyou.

Re: cmrunpkg : Script failed with no restart

John Bigg — Thu, 08 Dec 2005 07:44:20 GMT

Are you saying that the file /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl.log is empty on both nodes? That would be very strange.

I suggest you look at the script itself to look to see what it should be doing. I suspect the cause of the problem also results in the message:

Dec 8 10:41:38 sglx02 modprobe: modprobe: Can't locate module block-major-43

I suggets you look at the script to see what it is doing with kernel modules and look to see if this error is relevant and might be related to the cause.

Re: cmrunpkg : Script failed with no restart

Stephen Doud — Thu, 08 Dec 2005 09:41:46 GMT

Try manually start (and then stop) the control script
# /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl start
--- if it generates messages, then the log file in the same directory should have been created and appended when Serviceguard starts the package via cmrunpkg (after the package is -enabled- to start.

Check the package enablement using: cmviewcl -v -p
To set AUTO_RUN (general authorization to run the package on any adoptive node), use:
# cmmodpkg -e
To set NODE_Switching (per-node authorization to run the package), use:
# cmmodpkg -e -n

Re: cmrunpkg : Script failed with no restart

NiCK_76 — Thu, 08 Dec 2005 10:59:13 GMT

hi

I manual started http_pkg.cntl and checked all config file has parameter SERVICE_NAME.

[root@sglx02 conf]# /usr/local/cmcluster/conf/http_pkg1/http_pkg.cntl start

###### Node "sglx02": Starting package at Thu Dec 8 23:53:02 CST 2005 ######
Dec 8 23:53:02 - Node "sglx02": Activating volume group vgpkgB .
Dec 8 23:53:03 - Node "sglx02": Checking filesystems:
/dev/vgpkgB/lvol1
e2fsck 1.32 (09-Nov-2002)
/dev/vgpkgB/lvol1: clean, 41/77920 files, 2495/155648 blocks
Dec 8 23:53:03 - Node "sglx02": Mounting /dev/vgpkgB/lvol1 at /home2
Dec 8 23:53:03 - Node "sglx02": Adding IP address 192.168.15.119 to subnet 192.168.8.0
Dec 8 23:53:03 - Node "sglx02": Starting Apache daemons.
Dec 8 23:53:05 - Node "sglx02": Starting httpd.monitor using /usr/local/cmcluster/conf/http_pkg1/hahttp.mon
cmrunserv : Service name httpd.monitor does not exist.
ERROR: Function start_monitor_service: Failed to start monitor process.
Dec 8 23:53:05 - Node "sglx02": Halting the Apache service httpd.monitor
cmhaltserv : Service name httpd.monitor is not running.
Dec 8 23:53:05 - Node "sglx02": Stopping Apache daemons
ERROR: Dec 8 23:53:10 - Failed to start HTTP.
ERROR: Function verify_ha_server; Failed to start HA servers
Dec 8 23:53:10 - Node "sglx02": Remove IP address 192.168.15.119 from subnet 192.168.8.0
Dec 8 23:53:10 - Node "sglx02": Unmounting filesystem on /home2
Dec 8 23:53:10 - Node "sglx02": Deactivating volume group vgpkgB
###### Node "sglx02": Package start FAILED at Thu Dec 8 23:53:10 CST 2005 ######

Re: cmrunpkg : Script failed with no restart

John Bigg — Thu, 08 Dec 2005 11:30:16 GMT

Unfortunately, the error you see here associated with the service is as a result of running the package script manually. You really need this log information from the package log file after running cmrunpkg.

Re: cmrunpkg : Script failed with no restart

NiCK_76 — Fri, 09 Dec 2005 10:19:35 GMT

Hi all,

I reinstall the apache toolkit what according to the README file in the toolkit rpm very carefully. How ever, the problem still same.

Re: cmrunpkg : Script failed with no restart

melvyn burnard — Sat, 10 Dec 2005 12:48:22 GMT

Now is the time to log a call with your local HP Repsonse Centre

Re: cmrunpkg : Script failed with no restart

John Bigg — Tue, 13 Dec 2005 05:47:15 GMT

Whoever investigates this problem either here or in the response centre will need the information logged into the package log file when the package fails to start. This is the only way to determine what the problem is. If you can provide this then we can help. Without it it will be impossible to determine what the cause is. If the package log file really is not being created then this needs to be troubleshooted first. You will probably require help from the response centre to do this since it will almost certainly involve turning up logging on the daemons.