Hi all,

Thanks for your advice, I have double checked mod_status in Apache and its
definitely enabled on both nodes :
ls /etc/apache2/mods-enabled
alias.conf            authz_user.load  dir.conf          reqtimeout.conf
alias.load            autoindex.conf   dir.load          reqtimeout.load
auth_basic.load       autoindex.load   env.load          setenvif.conf
authn_file.load       cgid.conf        mime.conf         setenvif.load
authz_default.load    cgid.load        mime.load         status.conf
authz_groupfile.load  deflate.conf     negotiation.conf  status.load
authz_host.load       deflate.load     negotiation.load

I have checked the status page http://node/server-status and I can see the
status page ok. The mod_status is enabled on my node and runs fine.

I had a look at my apache log as you advised but I can't see Apache moaning
about a specific error, apart from multiple stops and restarts due to my
tests :

[Thu Sep 15 14:20:37 2011] [notice] caught SIGTERM, shutting down
[Thu Sep 15 14:20:38 2011] [notice] Apache/2.2.17 (Ubuntu) configured --
resuming normal operations
[Thu Sep 15 14:20:38 2011] [error] (2)No such file or directory: Couldn't
bind unix domain socket /var/run/apache2/cgisock.4278
[Thu Sep 15 14:20:39 2011] [notice] caught SIGTERM, shutting down

That's for the primary node. It looks like Corosync shuts down Apache.
On the Apache log file of the second node, I see the following :

[Thu Sep 15 14:18:27 2011] [notice] Apache/2.2.17 (Ubuntu) configured --
resuming normal operations
[Thu Sep 15 14:18:27 2011] [error] (2)No such file or directory: Couldn't
bind unix domain socket /var/run/apache2/cgisock.1338
[Thu Sep 15 14:18:28 2011] [crit] cgid daemon failed to initialize

I still have errors but the http service keep on running, no SIGTERM.

And then my node status is :

Online: [ node1 node2 ]

 Resource Group: group1
     failover-ip (ocf::heartbeat:IPaddr): Started node1
     apache (ocf::heartbeat:apache): Stopped

Failed actions:
    apache_start_0 (node=node2, call=6, rc=1, status=complete): unknown
error
    apache_monitor_0 (node=node1, call=3, rc=1, status=complete): unknown
error
    apache_start_0 (node=node1, call=7, rc=1, status=complete): unknown
error


of interest, some information found in var/log/syslog on node1 :

Sep 15 02:32:56 node1 crmd: [710]: info: do_state_transition: All 2 cluster
nodes are eligible to run resources.
Sep 15 02:32:56 node1 apache[928]: INFO: apache not running
Sep 15 02:32:56 node1 apache[928]: INFO: waiting for apache
/etc/apache2/apache2.conf to come up

Sep 15 02:32:58 node1 apache[928]: INFO: Killing apache PID 995
Sep 15 02:32:59 node1 lrmd: [707]: info: RA output: (apache:start:stderr)
kill: 833:
Sep 15 02:32:59 node1 lrmd: [707]: info: RA output: (apache:start:stderr) No
such process
Sep 15 02:32:59 node1 lrmd: [707]: info: RA output: (apache:start:stderr)
Sep 15 02:32:59 node1 apache[928]: INFO: Killing apache PID 995
Sep 15 02:32:59 node1 apache[928]: INFO: apache stopped.
Sep 15 02:32:59 node1 crmd: [710]: info: process_lrm_event: LRM operation
apache_start_0 (call=6, rc=1, cib-update=37, confirmed=true) unknown error
Sep 15 02:32:59 node1 crmd: [710]: WARN: status_from_rc: Action 8
(apache_start_0) on node1 failed (target: 0 vs. rc: 1): Error
Sep 15 02:32:59 node1 crmd: [710]: WARN: update_failcount: Updating
failcount for apache on node1 after failed start: rc=1 (update=INFINITY,
time=1316050379)
Sep 15 02:32:59 node1 crmd: [710]: info: abort_transition_graph:
match_graph_event:272 - Triggered transition abort (complete=0,
tag=lrm_rsc_op, id=apache_start_0,
magic=0:1;8:3:0:a4e41810-3e8f-439a-9b92-489edf657291, cib=0.172.10) : Event
failed
Sep 15 02:32:59 node1 crmd: [710]: info: update_abort_priority: Abort
priority upgraded from 0 to 1
Sep 15 02:32:59 node1 crmd: [710]: info: update_abort_priority: Abort action
done superceeded by restart
Sep 15 02:32:59 node1 crmd: [710]: info: match_graph_event: Action
apache_start_0 (8) confirmed on node1 (rc=4)
Sep 15 02:32:59 node1 crmd: [710]: info: run_graph:
====================================================
Sep 15 02:32:59 node1 crmd: [710]: notice: run_graph: Transition 3
(Complete=3, Pending=0, Fired=0, Skipped=4, Incomplete=0,
Source=/var/lib/pengine/pe-input-247.bz2): Stopped
Sep 15 02:32:59 node1 crmd: [710]: info: te_graph_trigger: Transition 3 is
now complete
Sep 15 02:32:59 node1 crmd: [710]: info: do_state_transition: State
transition S_TRANSITION_ENGINE -> S_POLICY_ENGINE [ input=I_PE_CALC
cause=C_FSA_INTERNAL origin=notify_crmd ]
Sep 15 02:32:59 node1 crmd: [710]: info: do_state_transition: All 2 cluster
nodes are eligible to run resources.




Here is my crm configure show, is there anything I can change there ?


root@node1:/home/user# crm configure show
node node1 \
attributes standby="off"
node node2 \
attributes standby="off"
primitive apache ocf:heartbeat:apache \
params configfile="/etc/apache2/apache2.conf" httpd="/usr/sbin/apache2" \
op start interval="10" timeout="40s" \
 op stop interval="10" timeout="60s" \
op monitor interval="5s"
primitive failover-ip ocf:heartbeat:IPaddr \
params ip="192.168.0.105" \
op monitor interval="5s"
group group1 failover-ip apache
location cli-prefer-failover-ip failover-ip \
rule $id="cli-prefer-rule-failover-ip" inf: #uname eq node1
property $id="cib-bootstrap-options" \
dc-version="1.0.9-da7075976b5ff0bee71074385f8fd02f296ec8a3" \
cluster-infrastructure="openais" \
 expected-quorum-votes="2" \
stonith-enabled="false" \
 no-quorum-policy="ignore"


Thank you for your help,

Guillaume


On 13 September 2011 08:29, Tim Serong <[email protected]> wrote:

> On 13/09/11 00:39, Guillaume Bettayeb wrote:
> > Hi there,
> >
> > This is my first post on this list, so hello everybody :)
> >
> > I am currently testing the fun of Linux HA Clustering (just for
> > personal interest)
> > and I have successfully set up a tiny ubuntu virtualbox 2 nodes
> > cluster with Ip failover and Apache running as resources.
> >
> > Right after the install, I tried to move the resources from a node to
> > the other (command standby) an everything worked like a charm.
> > Then I tried some failure tests, and started with a simple
> > /etc/init.d/networking stop on one node, noticed that the other node
> > took ownership of the resources automatically, all was fine.
> >
> > Then I have rebooted the nodes just to see how they would restart the
> > cluster, and since I have the following error :
> >
> > apache_start_0 (node=node1, call=8, rc=1, status=complete): unknown error
> >
> > For reading convenience,  my outputs are available at
> > http://pastebin.com/w1J4TWaG
> > Just to clarify, that's :
> > - crm configure show command
> > - crm_mon status
> > - All relevant information into my /var/log/syslog (although I was not
> > sure what to look at, I never used corosync before)
> >
> > I have read on an older post that the apache error usually have
> > something to do with either timeout or mod status.
> > As you can see on my pastebin, my timeout values are ok :
> > op stop interval="60s" timeout="120" \
> > op start interval="60s" timeout="120" \
> >
> > as for mod_status it's already enabled in Apache :
> > root@node1:/etc/apache2# a2enmod status
> > Module status already enabled
> >
> > Have I done anything wrong or is there anything else I should
> check/configure ?
> >
> > Any help with this matter would be greatly appreciated :)
>
> On a punt, it's probably mod_status.  Check your Apache logs at the time
> the start failed.  If it's whining about a 403 or 404 for /server-status
> (or similar), you need to fix that in your Apache config.
>
> HTH,
>
> Tim
> --
> Tim Serong
> Senior Clustering Engineer
> SUSE
> [email protected]
> _______________________________________________
> Linux-HA mailing list
> [email protected]
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems
>
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Reply via email to