microsoft.public.sqlserver.server - SQLServerAgent not starting

Author

SQLServerAgent not starting

Mark Allison

2002-08-14, 4:23 am

Hi,

Using SQL7SP3 on Win NT4 Server SP6a.

We did a disaster recovery test last week and switched off
our main site. We failed over to the DR site where we had
a BDC providing authentication. The PDC was unavailable.
After much bottom-clenching and nail-biting, we found that
SQLServerAgent would not start on one of our SQL Servers.
When clicking start in Control Panel-->Services, we got
this error:

"Error 2186: The service is not responding to the control
function."

I cleared the mail profile and stopped sql server and
sqlagent and I got the same message. I rebooted the server
a couple of times to no avail. We then needed to open a
pinhole in the firewall so that the PDC was visiable and
SQLServerAgent started. I'm not sure if this is a
coincidence or whether it's related. Has anyone else got
an explanation? The SQLAgent service used the same domain
account with local admin rights as the MSSQLServer service
and I could start and stop MSSQLServer with no problems.

The PDC was since removed from the network and SQLAgent
continued to run - however I did not stop and restart it
to test whether the statup process was to blame. All our
other SQL Servers did not experience this problem.

Can anyone shed any light on this?

Thanks,
Mark.

____________________________
Mark Allison, SQL Server DBA
Tel: +44 (0) 207 971 5894

Andrew J. Kelly

2002-08-14, 7:23 am

Sounds to me like the BDC was never synchronized with the PDC and the
account you use for the SQL Agent is just not valid according to the BDC.
One it's started it doesn't have to authenticate again for normal operation
so I suspect it was validated thru the PDC when you opened the Firewall. I
would ensure the BDC is totally synced with the PDC.

--
Andrew J. Kelly SQL MVP
Targitinteractive, Inc.

"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
news:1eb601c2437a$2759e8b0$37e
f2ecf@TKMSFTNGXA13...
> Hi,
>
> Using SQL7SP3 on Win NT4 Server SP6a.
>
> We did a disaster recovery test last week and switched off
> our main site. We failed over to the DR site where we had
> a BDC providing authentication. The PDC was unavailable.
> After much bottom-clenching and nail-biting, we found that
> SQLServerAgent would not start on one of our SQL Servers.
> When clicking start in Control Panel-->Services, we got
> this error:
>
> "Error 2186: The service is not responding to the control
> function."
>
> I cleared the mail profile and stopped sql server and
> sqlagent and I got the same message. I rebooted the server
> a couple of times to no avail. We then needed to open a
> pinhole in the firewall so that the PDC was visiable and
> SQLServerAgent started. I'm not sure if this is a
> coincidence or whether it's related. Has anyone else got
> an explanation? The SQLAgent service used the same domain
> account with local admin rights as the MSSQLServer service
> and I could start and stop MSSQLServer with no problems.
>
> The PDC was since removed from the network and SQLAgent
> continued to run - however I did not stop and restart it
> to test whether the statup process was to blame. All our
> other SQL Servers did not experience this problem.
>
> Can anyone shed any light on this?
>
> Thanks,
> Mark.
>
> ____________________________
> Mark Allison, SQL Server DBA
> Tel: +44 (0) 207 971 5894
>

Mark Allison

2002-08-14, 7:23 am

Thanks for your reply, Andrew.

This cannot be the case because the MSSQLServer service is
using the same account and that can be started and stopped
with no problem. In any case the account that MSSQLServer
service and SQLServerAgent service uses hasn't changed in
months.

Thanks,
Mark.

>-----Original Message-----
>Sounds to me like the BDC was never synchronized with the
PDC and the
>account you use for the SQL Agent is just not valid
according to the BDC.
>One it's started it doesn't have to authenticate again
for normal operation
>so I suspect it was validated thru the PDC when you
opened the Firewall. I
>would ensure the BDC is totally synced with the PDC.
>
>--
>Andrew J. Kelly SQL MVP
>Targitinteractive, Inc.
>
>
>
>"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> news:1eb601c2437a$2759e8b0$37e
f2ecf@TKMSFTNGXA13...
>> Hi,
>>
>> Using SQL7SP3 on Win NT4 Server SP6a.
>>
>> We did a disaster recovery test last week and switched
off
>> our main site. We failed over to the DR site where we
had
>> a BDC providing authentication. The PDC was unavailable.
>> After much bottom-clenching and nail-biting, we found
that
>> SQLServerAgent would not start on one of our SQL
Servers.
>> When clicking start in Control Panel-->Services, we got
>> this error:
>>
>> "Error 2186: The service is not responding to the
control
>> function."
>>
>> I cleared the mail profile and stopped sql server and
>> sqlagent and I got the same message. I rebooted the
server
>> a couple of times to no avail. We then needed to open a
>> pinhole in the firewall so that the PDC was visiable and
>> SQLServerAgent started. I'm not sure if this is a
>> coincidence or whether it's related. Has anyone else got
>> an explanation? The SQLAgent service used the same
domain
>> account with local admin rights as the MSSQLServer
service
>> and I could start and stop MSSQLServer with no problems.
>>
>> The PDC was since removed from the network and SQLAgent
>> continued to run - however I did not stop and restart it
>> to test whether the statup process was to blame. All our
>> other SQL Servers did not experience this problem.
>>
>> Can anyone shed any light on this?
>>
>> Thanks,
>> Mark.
>>
>> ____________________________
>> Mark Allison, SQL Server DBA
>> Tel: +44 (0) 207 971 5894
>>
>
>
>.
>

Andrew J. Kelly

2002-08-14, 8:23 am

Sorry I did not catch that part. I have not seen that particular error but
this seems to be related:

http://support.microsoft.com/defaul...b;en-us;Q183194

--
Andrew J. Kelly SQL MVP
Targitinteractive, Inc.

"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
news:23a401c24392$9e9b2890$b1e
62ecf@tkmsftngxa04...
> Thanks for your reply, Andrew.
>
> This cannot be the case because the MSSQLServer service is
> using the same account and that can be started and stopped
> with no problem. In any case the account that MSSQLServer
> service and SQLServerAgent service uses hasn't changed in
> months.
>
> Thanks,
> Mark.
>
> >-----Original Message-----
> >Sounds to me like the BDC was never synchronized with the
> PDC and the
> >account you use for the SQL Agent is just not valid
> according to the BDC.
> >One it's started it doesn't have to authenticate again
> for normal operation
> >so I suspect it was validated thru the PDC when you
> opened the Firewall. I
> >would ensure the BDC is totally synced with the PDC.
> >
> >--
> >Andrew J. Kelly SQL MVP
> >Targitinteractive, Inc.
> >
> >
> >
> >"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> > news:1eb601c2437a$2759e8b0$37e
f2ecf@TKMSFTNGXA13...
> >> Hi,
> >>
> >> Using SQL7SP3 on Win NT4 Server SP6a.
> >>
> >> We did a disaster recovery test last week and switched
> off
> >> our main site. We failed over to the DR site where we
> had
> >> a BDC providing authentication. The PDC was unavailable.
> >> After much bottom-clenching and nail-biting, we found
> that
> >> SQLServerAgent would not start on one of our SQL
> Servers.
> >> When clicking start in Control Panel-->Services, we got
> >> this error:
> >>
> >> "Error 2186: The service is not responding to the
> control
> >> function."
> >>
> >> I cleared the mail profile and stopped sql server and
> >> sqlagent and I got the same message. I rebooted the
> server
> >> a couple of times to no avail. We then needed to open a
> >> pinhole in the firewall so that the PDC was visiable and
> >> SQLServerAgent started. I'm not sure if this is a
> >> coincidence or whether it's related. Has anyone else got
> >> an explanation? The SQLAgent service used the same
> domain
> >> account with local admin rights as the MSSQLServer
> service
> >> and I could start and stop MSSQLServer with no problems.
> >>
> >> The PDC was since removed from the network and SQLAgent
> >> continued to run - however I did not stop and restart it
> >> to test whether the statup process was to blame. All our
> >> other SQL Servers did not experience this problem.
> >>
> >> Can anyone shed any light on this?
> >>
> >> Thanks,
> >> Mark.
> >>
> >> ____________________________
> >> Mark Allison, SQL Server DBA
> >> Tel: +44 (0) 207 971 5894
> >>
> >
> >
> >.
> >

Tibor Karaszi

2002-08-14, 8:23 am

Also, see if there is anything in Agents error log file.

--
Tibor Karaszi, SQL Server MVP
Archive at: http://groups.google.com/groups?oi=...ublic.sqlserver

"Andrew J. Kelly" <akelly@targitinteractive.com> wrote in message
news:uWW7ae5QCHA.3772@tkmsftngp08...
> Sorry I did not catch that part. I have not seen that particular error but
> this seems to be related:
>
> http://support.microsoft.com/defaul...b;en-us;Q183194
>
> --
> Andrew J. Kelly SQL MVP
> Targitinteractive, Inc.
>
>
>
> "Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> news:23a401c24392$9e9b2890$b1e
62ecf@tkmsftngxa04...
> > Thanks for your reply, Andrew.
> >
> > This cannot be the case because the MSSQLServer service is
> > using the same account and that can be started and stopped
> > with no problem. In any case the account that MSSQLServer
> > service and SQLServerAgent service uses hasn't changed in
> > months.
> >
> > Thanks,
> > Mark.
> >
> > >-----Original Message-----
> > >Sounds to me like the BDC was never synchronized with the
> > PDC and the
> > >account you use for the SQL Agent is just not valid
> > according to the BDC.
> > >One it's started it doesn't have to authenticate again
> > for normal operation
> > >so I suspect it was validated thru the PDC when you
> > opened the Firewall. I
> > >would ensure the BDC is totally synced with the PDC.
> > >
> > >--
> > >Andrew J. Kelly SQL MVP
> > >Targitinteractive, Inc.
> > >
> > >
> > >
> > >"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> > > news:1eb601c2437a$2759e8b0$37e
f2ecf@TKMSFTNGXA13...
> > >> Hi,
> > >>
> > >> Using SQL7SP3 on Win NT4 Server SP6a.
> > >>
> > >> We did a disaster recovery test last week and switched
> > off
> > >> our main site. We failed over to the DR site where we
> > had
> > >> a BDC providing authentication. The PDC was unavailable.
> > >> After much bottom-clenching and nail-biting, we found
> > that
> > >> SQLServerAgent would not start on one of our SQL
> > Servers.
> > >> When clicking start in Control Panel-->Services, we got
> > >> this error:
> > >>
> > >> "Error 2186: The service is not responding to the
> > control
> > >> function."
> > >>
> > >> I cleared the mail profile and stopped sql server and
> > >> sqlagent and I got the same message. I rebooted the
> > server
> > >> a couple of times to no avail. We then needed to open a
> > >> pinhole in the firewall so that the PDC was visiable and
> > >> SQLServerAgent started. I'm not sure if this is a
> > >> coincidence or whether it's related. Has anyone else got
> > >> an explanation? The SQLAgent service used the same
> > domain
> > >> account with local admin rights as the MSSQLServer
> > service
> > >> and I could start and stop MSSQLServer with no problems.
> > >>
> > >> The PDC was since removed from the network and SQLAgent
> > >> continued to run - however I did not stop and restart it
> > >> to test whether the statup process was to blame. All our
> > >> other SQL Servers did not experience this problem.
> > >>
> > >> Can anyone shed any light on this?
> > >>
> > >> Thanks,
> > >> Mark.
> > >>
> > >> ____________________________
> > >> Mark Allison, SQL Server DBA
> > >> Tel: +44 (0) 207 971 5894
> > >>
> > >
> > >
> > >.
> > >
>
>

Mark Allison

2002-08-14, 9:23 am

Tibor,

Nothing of note in the SQLAgent log file. I think I am
going to have to raise an MS PSS call. The only problem is
I can't reproduce it. The service is working OK now on our
primary site.

Management want confirmation that it's not going to happen
again if we do another DR test. I am in a sticky situation
becuause

a) I cannot confirm why it didn't work in the first place
b) I cannot give reassurance that it won't happen again
c) I cannot reproduce the problem for MS PSS to work on.

:-/

Mark.
>-----Original Message-----
>Also, see if there is anything in Agents error log file.
>
>--
>Tibor Karaszi, SQL Server MVP
>Archive at: http://groups.google.com/groups?
oi=djq&as_ugroup=microsoft.public.sqlserver
>
>

Mark Allison

2002-08-14, 9:23 am

Thanks, I found that but it is not related to my problem.

Mark.

>-----Original Message-----
>Sorry I did not catch that part. I have not seen that
particular error but
>this seems to be related:
>
>http://support.microsoft.com/default.aspx?scid=kb;en-
us;Q183194
>
>--
>Andrew J. Kelly SQL MVP
>Targitinteractive, Inc.
>
>
>
>"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> news:23a401c24392$9e9b2890$b1e
62ecf@tkmsftngxa04...
>> Thanks for your reply, Andrew.
>>
>> This cannot be the case because the MSSQLServer service
is
>> using the same account and that can be started and
stopped
>> with no problem. In any case the account that
MSSQLServer
>> service and SQLServerAgent service uses hasn't changed
in
>> months.
>>
>> Thanks,
>> Mark.
>>
>> >-----Original Message-----
>> >Sounds to me like the BDC was never synchronized with
the
>> PDC and the
>> >account you use for the SQL Agent is just not valid
>> according to the BDC.
>> >One it's started it doesn't have to authenticate again
>> for normal operation
>> >so I suspect it was validated thru the PDC when you
>> opened the Firewall. I
>> >would ensure the BDC is totally synced with the PDC.
>> >
>> >--
>> >Andrew J. Kelly SQL MVP
>> >Targitinteractive, Inc.
>> >
>> >
>> >
>> >"Mark Allison" <mark@allisonmitchell.c0m> wrote in
message
>> > news:1eb601c2437a$2759e8b0$37e
f2ecf@TKMSFTNGXA13...
>> >> Hi,
>> >>
>> >> Using SQL7SP3 on Win NT4 Server SP6a.
>> >>
>> >> We did a disaster recovery test last week and
switched
>> off
>> >> our main site. We failed over to the DR site where we
>> had
>> >> a BDC providing authentication. The PDC was
unavailable. [colo
r=darkred]
>> >> After much bottom-clenching and nail-biting, we found
>> that
>> >> SQLServerAgent would not start on one of our SQL
>> Servers.
>> >> When clicking start in Control Panel-->Services, we[/color]
got[c
olor=darkred]
>> >> this error:
>> >>
>> >> "Error 2186: The service is not responding to the
>> control
>> >> function."
>> >>
>> >> I cleared the mail profile and stopped sql server and
>> >> sqlagent and I got the same message. I rebooted the
>> server
>> >> a couple of times to no avail. We then needed to[/color]
open a[col
or=darkred]
>> >> pinhole in the firewall so that the PDC was visiable[/color]
and[c
olor=darkred]
>> >> SQLServerAgent started. I'm not sure if this is a
>> >> coincidence or whether it's related. Has anyone else[/color]
got[c
olor=darkred]
>> >> an explanation? The SQLAgent service used the same
>> domain
>> >> account with local admin rights as the MSSQLServer
>> service
>> >> and I could start and stop MSSQLServer with no[/color]
problems. [colo
r=darkred]
>> >>
>> >> The PDC was since removed from the network and[/color]
SQLAgent
>> >> continued to run - however I did not stop and
restart it[co
lor=darkred]
>> >> to test whether the statup process was to blame. All[/color]
our[c
olor=darkred]
>> >> other SQL Servers did not experience this problem.
>> >>
>> >> Can anyone shed any light on this?
>> >>
>> >> Thanks,
>> >> Mark.
>> >>
>> >> ____________________________
>> >> Mark Allison, SQL Server DBA
>> >> Tel: +44 (0) 207 971 5894
>> >>
>> >
>> >
>> >.
>> >
>
>
>.
>[/color]

Tibor Karaszi

2002-08-14, 9:23 am

Ouch. The worst kind... :-(

--
Tibor Karaszi, SQL Server MVP
Archive at: http://groups.google.com/groups?oi=...ublic.sqlserver

"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
news:1fea01c243a4$05d248c0$37e
f2ecf@TKMSFTNGXA13...
> Tibor,
>
> Nothing of note in the SQLAgent log file. I think I am
> going to have to raise an MS PSS call. The only problem is
> I can't reproduce it. The service is working OK now on our
> primary site.
>
> Management want confirmation that it's not going to happen
> again if we do another DR test. I am in a sticky situation
> becuause
>
> a) I cannot confirm why it didn't work in the first place
> b) I cannot give reassurance that it won't happen again
> c) I cannot reproduce the problem for MS PSS to work on.
>
> :-/
>
> Mark.
> >-----Original Message-----
> >Also, see if there is anything in Agents error log file.
> >
> >--
> >Tibor Karaszi, SQL Server MVP
> >Archive at: http://groups.google.com/groups?
> oi=djq&as_ugroup=microsoft.public.sqlserver
> >
> >

Mark Allison

2002-08-14, 10:23 am

OK, I'm going to reapply NT4 Service Pack 6a and SQL7 SP3
and hope that it works. Management have asked for another
site failover next week as they were not happy it didn't
work to plan last time.

If it doesn't work, I will be looking for another job... :-
)

Mark.

>-----Original Message-----
>Ouch. The worst kind... :-(
>
>--
>Tibor Karaszi, SQL Server MVP
>Archive at: http://groups.google.com/groups?
oi=djq&as_ugroup=microsoft.public.sqlserver
>
>
>"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> news:1fea01c243a4$05d248c0$37e
f2ecf@TKMSFTNGXA13...
>> Tibor,
>>
>> Nothing of note in the SQLAgent log file. I think I am
>> going to have to raise an MS PSS call. The only problem
is
>> I can't reproduce it. The service is working OK now on
our
>> primary site.
>>
>> Management want confirmation that it's not going to
happen
>> again if we do another DR test. I am in a sticky
situation
>> becuause
>>
>> a) I cannot confirm why it didn't work in the first
place

>> b) I cannot give reassurance that it won't happen again
>> c) I cannot reproduce the problem for MS PSS to work on.
>>
>> :-/
>>
>> Mark.
>> >-----Original Message-----
>> >Also, see if there is anything in Agents error log
file. [colo
r=darkred]
>> >
>> >--
>> >Tibor Karaszi, SQL Server MVP
>> >Archive at: http://groups.google.com/groups?
>> oi=djq&as_ugroup=microsoft.public.sqlserver
>> >
>> >
>
>
>.
>[/color]

Jasper Smith

2002-08-14, 1:23 pm

Mark,
In a true DR situation does the plan call for the promotion
of the BDC to a PDC ? Is it simply simulating SQL server
failure or loss of the entire subnet that the servers are on.
Was the SQL Agent service able to start when switched to
Local System as startup account , was this attempted ?
It does seem odd that the SQL service started succesfully
when I'd expect it to use the same API calls to authenticate
the startup account as SQL Agent service.

Good Luck :-)

Cheers
Jasper

"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
news:1eed01c243a9$c578af70$36e
f2ecf@tkmsftngxa12...
> OK, I'm going to reapply NT4 Service Pack 6a and SQL7 SP3
> and hope that it works. Management have asked for another
> site failover next week as they were not happy it didn't
> work to plan last time.
>
> If it doesn't work, I will be looking for another job... :-
> )
>
> Mark.
>
> >-----Original Message-----
> >Ouch. The worst kind... :-(
> >
> >--
> >Tibor Karaszi, SQL Server MVP
> >Archive at: http://groups.google.com/groups?
> oi=djq&as_ugroup=microsoft.public.sqlserver
> >
> >
> >"Mark Allison" <mark@allisonmitchell.c0m> wrote in message
> > news:1fea01c243a4$05d248c0$37e
f2ecf@TKMSFTNGXA13...
> >> Tibor,
> >>
> >> Nothing of note in the SQLAgent log file. I think I am
> >> going to have to raise an MS PSS call. The only problem
> is

> >> I can't reproduce it. The service is working OK now on
> our
> >> primary site.
> >>
> >> Management want confirmation that it's not going to
> happen
> >> again if we do another DR test. I am in a sticky
> situation
> >> becuause
> >>
> >> a) I cannot confirm why it didn't work in the first
> place
> >> b) I cannot give reassurance that it won't happen again
> >> c) I cannot reproduce the problem for MS PSS to work on.
> >>
> >> :-/
> >>
> >> Mark.
> >> >-----Original Message-----
> >> >Also, see if there is anything in Agents error log
> file.
> >> >
> >> >--
> >> >Tibor Karaszi, SQL Server MVP
> >> >Archive at: http://groups.google.com/groups?
> >> oi=djq&as_ugroup=microsoft.public.sqlserver
> >> >
> >> >
> >
> >
> >.
> >

Mark Allison

2002-08-14, 2:23 pm

Jasper,

It depends. If the primary site is knocked out by bomb, flood, or fire then
yes we would promote the DR site to PDC. The site I am working at WAS bombed
five years ago by the IRA, so it looks like it's a target - that's why
management are so paranoid. And rightly so.

If we evacuate the building and decide to fail-over all systems to the DR
site, but the threat turns out to be a hoax, then we won't promote the BDC
as we will still have connectivity with the primary site. I didn't try using
LocalSystem account. I'll bear that in mind next time...

Thanks again for your attention.

Mark.

"Jasper Smith" <jasper_smith9@hotmail.com> wrote in message
news:eGKdcI8QCHA.4268@tkmsftngp08...
> Mark,
> In a true DR situation does the plan call for the promotion
> of the BDC to a PDC ? Is it simply simulating SQL server
> failure or loss of the entire subnet that the servers are on.
> Was the SQL Agent service able to start when switched to
> Local System as startup account , was this attempted ?
> It does seem odd that the SQL service started succesfully
> when I'd expect it to use the same API calls to authenticate
> the startup account as SQL Agent service.
>
> Good Luck :-)
>
> Cheers
> Jasper
>

John Bell

2002-08-14, 10:23 pm

"Mark Allison" <mark@allisonmitchell.c0m> wrote in message news:<Oi9Rml8QCHA.3936@tkmsftngp11>...
> Jasper,
>
> If we evacuate the building and decide to fail-over all systems to the DR
> site, but the threat turns out to be a hoax, then we won't promote the BDC
> as we will still have connectivity with the primary site. I didn't try using
> LocalSystem account. I'll bear that in mind next time...
>
> Thanks again for your attention.
>
> Mark.

I would have thought that you would always plan to promote the BDC if
you have to fail over because loss of communication with the PDC would
be the criteria to trigger this i.e. a pre-requisit for
restoring/restarting the database servers is that you have a PDC.

If you evacuate the building and it turns out to be a hoax, how do you
loose connectivity to the PDC? If you have to crash the servers on
evacuation then you should have a hot standby rather than a cold one
(IMO).

John

Sponsored Links