asked on

Cluster Services will not start

Specifically, we are trying to setup a two-node cluster to provide a highly available apache server. After reviewing the documentation, it appears that shared storage may not be necessary, though we would like to have the document root be on shared storage eventually.

We have followed the steps laid out in the howto:

http://www.redhat.com/docs/en-US/Red_Hat_Enterprise_Linux/5.2/html/Cluster_Administration/ap-httpd-service-CA.html

However when we try to start the service in Luci, the httpd service fails to start. We get the following errors in /var/log/messages:

Jun 29 16:22:36 habox1 clurgmgrd: [10855]: <err> Stopping Service
apache:Apache_Test_Srvr > Failed
Jun 29 16:22:36 habox1 clurgmgrd[10855]: <notice> stop on apache
"Apache_Test_Srvr" returned 1 (generic error)
Jun 29 16:22:36 habox1 clurgmgrd[10855]: <crit> #13: Service
service:Web_Server failed to stop cleanly
Jun 29 16:26:29 habox1 clurgmgrd[10855]: <notice> Starting disabled
service service:Web_Server
Jun 29 16:26:29 habox1 clurgmgrd: [10855]: <err> Looking For IP
Addresses [apache:Apache_Test_Srvr] > Failed - No IP Addresses Found
Jun 29 16:26:29 habox1 clurgmgrd[10855]: <notice> start on apache
"Apache_Test_Srvr" returned 1 (generic error)
Jun 29 16:26:29 habox1 clurgmgrd[10855]: <warning> #68: Failed to start
service:Web_Server; return value: 1
Jun 29 16:26:29 habox1 clurgmgrd[10855]: <notice> Stopping service
service:Web_Server
Jun 29 16:26:35 habox1 clurgmgrd: [10855]: <err> Checking Existence Of
File /var/run/cluster/apache/apache:Apache_Test_Srvr.pid
[apache:Apache_Test_Srvr] > Failed - File Doesn't Exist
Jun 29 16:26:35 habox1 clurgmgrd: [10855]: <err> Stopping Service
apache:Apache_Test_Srvr > Failed
Jun 29 16:26:35 habox1 clurgmgrd[10855]: <notice> stop on apache
"Apache_Test_Srvr" returned 1 (generic error)
Jun 29 16:26:35 habox1 clurgmgrd[10855]: <crit> #12: RG
service:Web_Server failed to stop; intervention required
Jun 29 16:26:35 habox1 clurgmgrd[10855]: <notice> Service
service:Web_Server is failed
Jun 29 16:26:35 habox1 clurgmgrd[10855]: <crit> #13: Service
service:Web_Server failed to stop cleanly

Can you advise us as to what the problem may be? Let us know if you need
more information.

my cluster.conf file created in web GUI (luci)
 
<?xml version="1.0"?>
<cluster alias="app_server" config_version="16" name="app_server">
        <fence_daemon clean_start="0" post_fail_delay="0" 
post_join_delay="3"/>
        <clusternodes>
                <clusternode name="habox2.nimh.nih.gov" nodeid="1" 
votes="1">
                        <fence>
                                <method name="1"/>
                        </fence>
                </clusternode>
                <clusternode name="habox1.nimh.nih.gov" nodeid="2" 
votes="1">
                        <fence>
                                <method name="1"/>
                        </fence>
                </clusternode>
        </clusternodes>
        <cman expected_votes="1" two_node="1"/>
        <fencedevices/>
        <rm>
                <failoverdomains/>
                <resources>
                        <apache config_file="conf/httpd.conf" 
name="Apache_Test_Srvr" server_root="/etc/httpd" shutdown_wait="0"/>
                        <ip address="172.16.52.151" monitor_link="1"/>
                        <script file="/etc/rc.d/init.d/httpd" 
name="script_test"/>
                        <fs device="/dev/hda2" force_fsck="0" 
force_unmount="0" fsid="36806" fstype="ext3" mountpoint="/HA" 
name="httpd_content" self_fence="1"/>
                </resources>
                <service autostart="1" exclusive="0" max_restarts="0" 
name="Web_Server" recovery="restart" restart_expire_time="0">
                        <apache ref="Apache_Test_Srvr">
                                <ip ref="172.16.52.151"/>
                                <script ref="script_test"/>
                                <fs ref="httpd_content"/>
                        </apache>
                </service>
        </rm>
</cluster>

Open in new window

Document1.pdf

Justin_Edmands

ASKER

need some help!

JabbaDow

I think it might be easier to use hearbeat with DRBD: www.linux-ha.org and www.drbd.org. They work very nicely together. Basically you set up some information in their config files about each other, which IP address they will share, etc. Then at the end of it all, you have the heartbeat process start up, which then starts up the DRBD shared storage and places symlinks on the system to point to the shared storage. Heartbeat then handles the starting and stopping of Apache. It is pretty easy to do, and both projects are very well documented and there are many how-to articles on the web for doing exactly what you want to do.

arnold

IMHO, it is better to load balance a web server rather than set it up in a fail over cluster.
You could use rsync to synchronize the document root data.

You could setup a cluster resource dealing with a specific IP.
This will deal with making the IP "available all the time"

One error I see is that you are not assigning an IP that will move with the web server.

You have to setup an IP that will move between/among the nodes.

Justin_Edmands

ASKER

already got DRBD to work and all. need to do RedHat Cluster Suite

ASKER CERTIFIED SOLUTION

JabbaDow

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial