Mailing List Archive: Linux-HA: Users

only SAPInstance doesn't start

Index | Next | Previous | View Threaded

ibnerazi at yahoo

Nov 9, 2009, 8:44 AM

Post #1 of 10 (1566 views)
Permalink

only SAPInstance doesn't start

I am testing a simplest configuration for a two node(active/passive) SAP cluster.
/oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.

Problem:
when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
SAPInstance resource never starts or fails to start. All other
resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
HA(hb_gui)

cib.xml is attached.

Details:
---------
# rpm -qa |grep heartbeat
heartbeat-2.1.4-0.16.2
heartbeat-cmpi-2.1.4-0.16.2
heartbeat-ldirectord-2.1.4-0.16.2
heartbeat-pils-2.1.4-0.16.2
heartbeat-stonith-2.1.4-0.16.2

I am configuring a two node active/passive SAP cluster.
ERP, SAP Databse, SAP Instance, all will run on the same machine/node.

Physical IP of node1: 192.168.41.236
Physical IP of node2: 192.168.41.238
Cluster/Virtual IP: 192.168.41.245

Physical hostname of node1: prdnode1
Physical hostname of node2: prdnode2
Virtual hostname: sapprd

/etc/hosts on both nodes is following
# cat /etc/hosts
prdnode1 192.168.41.236
prdnode2 192.168.41.238
sapprd 192.168.41.245

Installed the SAP/Database via the following command
# sapinst SAPINST_USE_HOSTNAME=sapprd

Installation completed successfully, and now I can start/stop SAP/Database from any node(prdnode1, prdnode2).
i.e 'lsnrctl start', 'startsap -vhost sapprd', 'stopsap -vhost sapprd', and 'lsnrctl stop' works quite fine/smooth

and when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
SAPInstance resource never starts or fails to start. All other
resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
HA(hb_gui)

please help me in this matter

Attachments:

cib.xml (5.26 KB)

dejanmm at fastmail

Nov 9, 2009, 11:08 AM

Post #2 of 10 (1473 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

On Mon, Nov 09, 2009 at 08:44:46AM -0800, Muhammad Sharfuddin wrote:
> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
>
> Problem:
> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
> SAPInstance resource never starts or fails to start. All other
> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
> HA(hb_gui)

Impossible to say without taking look at the logs. You can try
yourself: grep lrmd.*sap_instance /var/log/yourlog

Thanks,

Dejan
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

ibnerazi at yahoo

Nov 9, 2009, 11:36 AM

Post #3 of 10 (1507 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

--- On Mon, 11/9/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
>>
>> Problem:
>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
>> SAPInstance resource never starts or fails to start. All other
>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
>> HA(hb_gui)

>Impossible to say without taking look at the logs. You can try
>yourself: grep lrmd.*sap_instance /var/log/yourlog

# grep lrmd.*sap_instance /var/log/messages
Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start process (PID 20871) timed out (try 1). Killing with signal SIGTERM (15).
Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: operation start[15] on ocf::SAPInstance::sap_instance for client 13098, its parameters: InstanceName=[PRD_DVEBMGS01_sapprd] CRM_meta_timeout=[20000] crm_feature_set=[2.0] : pid [20871] timed out
Nov 7 20:44:45 prdnode1 lrmd: [13095]: info: rsc:sap_instance:16: stop
Nov 7 20:45:05 prdnode1 lrmd: [13095]: WARN: sap_instance:stop process (PID 21951) timed out (try 1). Killing with signal SIGTERM (15).
Nov 7 20:45:05 prdnode1 lrmd: [13095]: WARN: operation stop[16] on ocf::SAPInstance::sap_instance for client 13098, its parameters: InstanceName=[PRD_DVEBMGS01_sapprd] CRM_meta_timeout=[20000] crm_feature_set=[2.0] : pid [21951] timed out
Nov 7 21:01:51 prdnode1 lrmd: [13095]: info: rsc:sap_instance_ascs:35: monitor
Nov 7 21:02:42 prdnode1 lrmd: [13095]: info: rsc:sap_instance_dvebmgs:36: monitor
Nov 7 21:03:30 prdnode1 lrmd: [13095]: info: rsc:sap_instance_ascs:37: monitor
Nov 7 21:34:29 prdnode1 lrmd: [13095]: info: rsc:sap_instance:64: monitor
Nov 7 21:34:31 prdnode1 lrmd: [13095]: info: rsc:sap_instance:65: start
Nov 7 21:34:51 prdnode1 lrmd: [13095]: WARN: sap_instance:start process (PID 16037) timed out (try 1). Killing with signal SIGTERM (15).
Nov 7 21:34:51 prdnode1 lrmd: [13095]: WARN: operation start[65] on ocf::SAPInstance::sap_instance for client 13098, its parameters: InstanceName=[PRD_DVEBMGS01_sapprd] CRM_meta_timeout=[20000] crm_feature_set=[2.0] : pid [16037] timed out
Nov 7 21:34:53 prdnode1 lrmd: [13095]: info: rsc:sap_instance:66: stop
Nov 7 21:35:13 prdnode1 lrmd: [13095]: WARN: sap_instance:stop process (PID 16635) timed out (try 1). Killing with signal SIGTERM (15).
Nov 7 21:35:13 prdnode1 lrmd: [13095]: WARN: operation stop[66] on ocf::SAPInstance::sap_instance for client 13098, its parameters: InstanceName=[PRD_DVEBMGS01_sapprd] CRM_meta_timeout=[20000] crm_feature_set=[2.0] : pid [16635] timed out

Regards

_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

dejanmm at fastmail

Nov 10, 2009, 2:04 AM

Post #4 of 10 (1482 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

On Mon, Nov 09, 2009 at 11:36:54AM -0800, Muhammad Sharfuddin wrote:
> --- On Mon, 11/9/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
> >> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
> >> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
> >>
> >> Problem:
> >> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
> >> SAPInstance resource never starts or fails to start. All other
> >> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
> >> HA(hb_gui)
>
> >Impossible to say without taking look at the logs. You can try
> >yourself: grep lrmd.*sap_instance /var/log/yourlog
>
> # grep lrmd.*sap_instance /var/log/messages
> Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
> Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
> Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start process (PID 20871) timed out (try 1). Killing with signal SIGTERM (15).

The start operation timed out. Either increase the timeout (looks
like it's set to 20 seconds) or see why does it take so long.
SAP probably takes long.

Thanks,

Dejan
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

ibnerazi at yahoo

Nov 10, 2009, 4:00 AM

Post #5 of 10 (1471 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

--- On Tue, 11/10/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
> >>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
> >>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
> >>>
> >>> Problem:
> >>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
> >>> SAPInstance resource never starts or fails to start. All other
> >>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
> >>> HA(hb_gui)
>
> >> Impossible to say without taking look at the logs. You can try
> >> yourself: grep lrmd.*sap_instance /var/log/yourlog
>
> > # grep lrmd.*sap_instance /var/log/messages
> > Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
> > Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
> >
Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start
process (PID 20871) timed out (try 1). Killing with signal SIGTERM
(15).

> The start operation timed out. Either increase the timeout (looks
> like it's set to 20 seconds)
How, I dont know(very new to HA). any guide/url that teaches this
esp via hb_gui.

> or see why does it take so long.
> SAP probably takes long.
on this machine, SAP takes almost 70 seconds to start

Regards

_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

dejanmm at fastmail

Nov 10, 2009, 4:45 AM

Post #6 of 10 (1471 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

On Tue, Nov 10, 2009 at 04:00:14AM -0800, Muhammad Sharfuddin wrote:
>
> --- On Tue, 11/10/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
> > >>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
> > >>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
> > >>>
> > >>> Problem:
> > >>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
> > >>> SAPInstance resource never starts or fails to start. All other
> > >>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
> > >>> HA(hb_gui)
> >
> > >> Impossible to say without taking look at the logs. You can try
> > >> yourself: grep lrmd.*sap_instance /var/log/yourlog
> >
> > > # grep lrmd.*sap_instance /var/log/messages
> > > Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
> > > Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
> > >
> Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start
> process (PID 20871) timed out (try 1). Killing with signal SIGTERM
> (15).
>
> > The start operation timed out. Either increase the timeout (looks
> > like it's set to 20 seconds)
> How, I dont know(very new to HA). any guide/url that teaches this
> esp via hb_gui.

Not using the gui so can't give you the exact steps. You should
add a start operation to the resource and then set the timeout
attribute.

Thanks,

Dejan

> > or see why does it take so long.
> > SAP probably takes long.
> on this machine, SAP takes almost 70 seconds to start
>
> Regards
>
>
>
>
>
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Rolf.Schmidt at novell

Nov 10, 2009, 5:38 AM

Post #7 of 10 (1473 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

On Tue, 10 Nov 2009, Dejan Muhamedagic wrote:

> Hi,
>
> On Tue, Nov 10, 2009 at 04:00:14AM -0800, Muhammad Sharfuddin wrote:
>>
>> --- On Tue, 11/10/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
>>>>>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
>>>>>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
>>>>>>
>>>>>> Problem:
>>>>>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
>>>>>> SAPInstance resource never starts or fails to start. All other
>>>>>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
>>>>>> HA(hb_gui)
>>>
>>>>> Impossible to say without taking look at the logs. You can try
>>>>> yourself: grep lrmd.*sap_instance /var/log/yourlog
>>>
>>>> # grep lrmd.*sap_instance /var/log/messages
>>>> Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
>>>> Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
>>>>
>> Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start
>> process (PID 20871) timed out (try 1). Killing with signal SIGTERM
>> (15).
>>
>>> The start operation timed out. Either increase the timeout (looks
>>> like it's set to 20 seconds)
>> How, I dont know(very new to HA). any guide/url that teaches this
>> esp via hb_gui.
>
> Not using the gui so can't give you the exact steps. You should
> add a start operation to the resource and then set the timeout
> attribute.

Easy with the GUI, open the resource and go to the Tab with
operation. Add a start and a stop operation for the SAP Resources.
Use the default values, they will be 240/300 seconds and should
do the trick.

> Thanks,
>
> Dejan
>
>>> or see why does it take so long.
>>> SAP probably takes long.
>> on this machine, SAP takes almost 70 seconds to start
>>
>> Regards

Rolf Schmidt

(9 out of 10 voices in my head tell me I am NOT mad)

ibnerazi at yahoo

Nov 11, 2009, 4:10 AM

Post #8 of 10 (1447 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

--- On Tue, 11/10/09, Rolf Schmidt <Rolf.Schmidt [at] novell> wrote:
On Tue, 10 Nov 2009, Dejan Muhamedagic wrote:

>> Hi,
>>
>> On Tue, Nov 10, 2009 at 04:00:14AM -0800, Muhammad Sharfuddin wrote:
>>>
>>> --- On Tue, 11/10/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
>>>>>>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
>>>>>>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
>>>>>>>
>>>>>>> Problem:
>>>>>>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
>>>>>>> SAPInstance resource never starts or fails to start. All other
>>>>>>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
>>>>>>> HA(hb_gui)
>>>>
>>>>>> Impossible to say without taking look at the logs. You can try
>>>>>> yourself: grep lrmd.*sap_instance /var/log/yourlog
>>>>
>>>>> # grep lrmd.*sap_instance /var/log/messages
>>>>> Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
>>>>> Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
>>>>>
>>> Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start
>>> process (PID 20871) timed out (try 1). Killing with signal SIGTERM
>>> (15).
>>>
>>>> The start operation timed out. Either increase the timeout (looks
>>>> like it's set to 20 seconds)
>>> How, I dont know(very new to HA). any guide/url that teaches this
>>> esp via hb_gui.
>>
>> Not using the gui so can't give you the exact steps. You should
>> add a start operation to the resource and then set the timeout
>> attribute.

>
> Easy with the GUI, open the resource and go to the Tab with
> operation. Add a start and a stop operation for the SAP Resources.
> Use the default values, they will be 240/300 seconds and should
> do the trick.
I add the start/stop operations for SAP Resources with the default
timeout values, but still SAPInstance resource fails to start, and
all other resources starts successfully.

cib.xml, and /var/log/messages attached

# crm_verify -LV
crm_verify[28948]: 2009/09/28_22:37:15 WARN: unpack_rsc_op:
Processing failed op resource_SAP_Instance_start_0 on prdnode2:
unknown exec error
crm_verify[28948]: 2009/09/28_22:37:15 WARN:
common_apply_stickiness: Forcing resource_SAP_Instance away
from prdnode2 after 1000000 failures (max=1000000)
crm_verify[28948]: 2009/09/28_22:37:15 WARN: native_color: Resource resource_SAP_Instance cannot run anywhere
Warnings found during check: config may not be valid

Attachments:

cib.xml (6.14 KB)

messages (57.0 KB)

dejanmm at fastmail

Nov 11, 2009, 4:56 AM

Post #9 of 10 (1438 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

On Wed, Nov 11, 2009 at 04:10:12AM -0800, Muhammad Sharfuddin wrote:
> --- On Tue, 11/10/09, Rolf Schmidt <Rolf.Schmidt [at] novell> wrote:
> On Tue, 10 Nov 2009, Dejan Muhamedagic wrote:
>
> >> Hi,
> >>
> >> On Tue, Nov 10, 2009 at 04:00:14AM -0800, Muhammad Sharfuddin wrote:
> >>>
> >>> --- On Tue, 11/10/09, Dejan Muhamedagic <dejanmm [at] fastmail> wrote:
> >>>>>>> I am testing a simplest configuration for a two node(active/passive) SAP cluster.
> >>>>>>> /oracle, /sapmnt, /usr/sap, and /home/prdadm are on SAN with ext3 filesystems.
> >>>>>>>
> >>>>>>> Problem:
> >>>>>>> when I tries to run/start SAP from HA(hb_gui), it fails.. i.e
> >>>>>>> SAPInstance resource never starts or fails to start. All other
> >>>>>>> resources e.g IPaddr2, Filesystem, and SAPDatabase starts smoothly from
> >>>>>>> HA(hb_gui)
> >>>>
> >>>>>> Impossible to say without taking look at the logs. You can try
> >>>>>> yourself: grep lrmd.*sap_instance /var/log/yourlog
> >>>>
> >>>>> # grep lrmd.*sap_instance /var/log/messages
> >>>>> Nov 7 20:44:20 prdnode1 lrmd: [13095]: info: rsc:sap_instance:14: monitor
> >>>>> Nov 7 20:44:23 prdnode1 lrmd: [13095]: info: rsc:sap_instance:15: start
> >>>>>
> >>> Nov 7 20:44:43 prdnode1 lrmd: [13095]: WARN: sap_instance:start
> >>> process (PID 20871) timed out (try 1). Killing with signal SIGTERM
> >>> (15).
> >>>
> >>>> The start operation timed out. Either increase the timeout (looks
> >>>> like it's set to 20 seconds)
> >>> How, I dont know(very new to HA). any guide/url that teaches this
> >>> esp via hb_gui.
> >>
> >> Not using the gui so can't give you the exact steps. You should
> >> add a start operation to the resource and then set the timeout
> >> attribute.
>
> >
> > Easy with the GUI, open the resource and go to the Tab with
> > operation. Add a start and a stop operation for the SAP Resources.
> > Use the default values, they will be 240/300 seconds and should
> > do the trick.
> I add the start/stop operations for SAP Resources with the default
> timeout values, but still SAPInstance resource fails to start, and
> all other resources starts successfully.
>
> cib.xml, and /var/log/messages attached

Sep 28 22:36:19 prdnode2 crmd: [19676]: ERROR: process_lrm_event: LRM operation resource_SAP_Instance_start_0 (15) Timed Out (timeout=180000ms)

BTW, it would be really great if you would take some initiative
and try to read the logs yourself.

Thanks,

Dejan
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Rolf.Schmidt at novell

Nov 12, 2009, 2:39 AM

Post #10 of 10 (1442 views)
Permalink

Re: only SAPInstance doesn't start [In reply to]

Hi,

as Dejan wrote already, it is all in your logfiles:

Sep 28 22:36:19 prdnode2 lrmd: [19673]: WARN: operation start[15] on ocf::SAPInstance::resource_SAP_Instance for client 19676, its parameters: CRM_meta_enabled=[true] CRM_meta_start-delay=[0] InstanceName=[PRD_DVEBMGS01_sapprd] CRM_meta_role=[Started] CRM_meta_timeout=[180000] crm_feature_set=[2.0] CRM_meta_name=[start] : pid [24153] timed out

Sep 28 22:36:19 prdnode2 tengine: [21729]: WARN: update_failcount: Updating failcount for resource_SAP_Instance on 98de9eff-c0f9-4ce6-bc37-92e91fc57351 after failed start: rc=-2 (update=INFINITY)

so increase the values in

<op id="op_SAPInstanceStart" name="start" timeout="180" interval="0" start_delay="0" disabled="false" role="Started"/>
<op id="op_SAPInstanceStop" name="stop" timeout="240" interval="0" start_delay="0" disabled="false" role="Started"/>

for the timeout. As a rule of thumb, just double them and try again. Also cleanup the resource to get rid
of the failcount.

And as this is the default failing maybe you want to tell your SAP Consultant so he
can give feedback to SAP that the default was not enough in your case. Or maybe even
better, he might tell you what timeout you need for your SAP instance.

Rolf
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads