** James,

But when say the database server for some reason decides to reboot at 4.00 am on Sunday morning this will cause the remedy service to hang unless someone manually reboots it. Of course you can script your way around this but it would be so much nicer if the system itself would be smart enough to reconnect.

Regards,
Michiel


On 3/14/06, McKenzie, James J C-E LCMC HQISEC/L3 wrote:

**

Michiel:

I would not like this feature. Why? Because there is a retry limit in the ar.conf file. This forces the arserverd process on Unix and arserver.exe programs to terminate whenever the database fails. This is a safety feature that prevents certain actions from happening. However, this does not always happen and if users are logged in when the database fails can cause miscommunications. What I would like to see is a fix for a long standing problem and that is the error message generation process. If you cannot communicate with the database, then the error message should state so like this:

"The Remedy Action Request Server cannot communicate with its associated database. Please advise the Remedy Administrator of this difficulty" rather than the cryptic ARERR 91/92/93/94 stuff. This would provide more information than I need in order to fix the problem. This information and what the user was doing at the time will suffice.

If communications is lost with the database, a constantly bouncing ARS process/program would be more frustrating to my users than a dead system.

James McKenzie



-----Original Message-----
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG]On Behalf Of Michiel Beijen
Sent: Tuesday, March 14, 2006 6:34 AM
To: arslist@ARSLIST.ORG

Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?






** Of course it would be far better that the AR System service would try to reconnect to the database every x minutes when it can not connect to the database. But I guess this would be an enhancement request.

Shal we all submit this enhancement request?

Regards,
Michiel


On 3/14/06, Krishnakumar wrote:
Hi James,
Like Brian has mentioned, the ideal thing to do would be to
monitor the DB and stop the ARS service if the DB is down.

However, if this turns out to be a bit out of reach, you could atleast
make sure that ARS is stopped before the DB is brought up again. This
would ensure that ARS is not using the old connection to the DB. ARS can
then be started once the SQL Listeners are up and running. This should
ensure that ARS works fine.


Regards,
Krishnakumar

On Tue, 14 Mar 2006 05:44:21 +0530, Luke, Brian < Brian.Luke@SCHWAB.COM>
wrote:

> James,
> We ran in to this situation as well. Our specific config was
> running Windows app server and Linux DB server.
>
> The best thing to do is to stop ARS on the remote machine (from the DB
> machine) as part of the script stopping the database. Likewise starting
> ARS just after starting the DB. If they're on the same platform it's a
> bit simpler. If it's not on the same platform, you'll need to figure
> out how to issue commands remotely across platforms.
>
> What platforms are your respective systems on?
> Do you have Tivoli, NetIQ, or any other monitoring systems in place the
> systems?
>
> In our case, we installed a cygwin/SSH server on the Windows app server,
> and can issue commands from the Linux box that way.
>
> Another approach that works is to create a monitor to test the return
> code from TNSPING on the app server to your DB Instance. When it fails,
> stop ARS. When it works (and didn't last time it checked), start ARS.
>
> Hope this helps,
> -Brian
>
> -----Original Message-----
> From: Action Request System discussion list(ARSList)
> [mailto:arslist@ARSLIST.ORG] On Behalf Of James Pifer
> Sent: Monday, March 13, 2006 1:43 PM
> To: arslist@ARSLIST.ORG
> Subject: Should Remedy reconnect to the database if the connection is
> lost and then restored?
>
> We're running ARS 6.x with Helpdesk 4.0.3 on Windows 2000 Server. Our DB
> is Oracle 9i running on another server. We've had problems quite a few
> times on the database server where it either has rebooted or been taken
> down for maintenance. In these instances I have to completely restart
> the Remedy server. I can't even just stop and start the service.
>
> Is this normal or should the Remedy service reconnect to the database
> once it is back up?
>
> Thanks,
> James
>
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org


20060125This posting was submitted with HTML in it

20060125This posting was submitted with HTML in it


20060125This posting was submitted with HTML in it

**
Michiel:

You are correct. However, this can be and is done through scripting. What you need to 'cover' is when either the database is taken down out of schedule or communications is lost between the ARS system and the database system.

James McKenzie


-----Original Message-----
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG]On Behalf Of Michiel Beijen
Sent: Tuesday, March 14, 2006 7:18 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?


** James,

But when say the database server for some reason decides to reboot at 4.00 am on Sunday morning this will cause the remedy service to hang unless someone manually reboots it. Of course you can script your way around this but it would be so much nicer if the system itself would be smart enough to reconnect.

Regards,
Michiel


On 3/14/06, McKenzie, James J C-E LCMC HQISEC/L3 wrote:

**

Michiel:

I would not like this feature. Why? Because there is a retry limit in the ar.conf file. This forces the arserverd process on Unix and arserver.exe programs to terminate whenever the database fails. This is a safety feature that prevents certain actions from happening. However, this does not always happen and if users are logged in when the database fails can cause miscommunications. What I would like to see is a fix for a long standing problem and that is the error message generation process. If you cannot communicate with the database, then the error message should state so like this:

"The Remedy Action Request Server cannot communicate with its associated database. Please advise the Remedy Administrator of this difficulty" rather than the cryptic ARERR 91/92/93/94 stuff. This would provide more information than I need in order to fix the problem. This information and what the user was doing at the time will suffice.

If communications is lost with the database, a constantly bouncing ARS process/program would be more frustrating to my users than a dead system.

James McKenzie



-----Original Message-----
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG]On Behalf Of Michiel Beijen
Sent: Tuesday, March 14, 2006 6:34 AM
To: arslist@ARSLIST.ORG

Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?






** Of course it would be far better that the AR System service would try to reconnect to the database every x minutes when it can not connect to the database. But I guess this would be an enhancement request.

Shal we all submit this enhancement request?

Regards,
Michiel


On 3/14/06, Krishnakumar wrote:
Hi James,
Like Brian has mentioned, the ideal thing to do would be to
monitor the DB and stop the ARS service if the DB is down.

However, if this turns out to be a bit out of reach, you could atleast
make sure that ARS is stopped before the DB is brought up again. This
would ensure that ARS is not using the old connection to the DB. ARS can
then be started once the SQL Listeners are up and running. This should
ensure that ARS works fine.


Regards,
Krishnakumar

On Tue, 14 Mar 2006 05:44:21 +0530, Luke, Brian < Brian.Luke@SCHWAB.COM>
wrote:

> James,
> We ran in to this situation as well. Our specific config was
> running Windows app server and Linux DB server.
>
> The best thing to do is to stop ARS on the remote machine (from the DB
> machine) as part of the script stopping the database. Likewise starting
> ARS just after starting the DB. If they're on the same platform it's a
> bit simpler. If it's not on the same platform, you'll need to figure
> out how to issue commands remotely across platforms.
>
> What platforms are your respective systems on?
> Do you have Tivoli, NetIQ, or any other monitoring systems in place the
> systems?
>
> In our case, we installed a cygwin/SSH server on the Windows app server,
> and can issue commands from the Linux box that way.
>
> Another approach that works is to create a monitor to test the return
> code from TNSPING on the app server to your DB Instance. When it fails,
> stop ARS. When it works (and didn't last time it checked), start ARS.
>
> Hope this helps,
> -Brian
>
> -----Original Message-----
> From: Action Request System discussion list(ARSList)
> [mailto:arslist@ARSLIST.ORG] On Behalf Of James Pifer
> Sent: Monday, March 13, 2006 1:43 PM
> To: arslist@ARSLIST.ORG
> Subject: Should Remedy reconnect to the database if the connection is
> lost and then restored?
>
> We're running ARS 6.x with Helpdesk 4.0.3 on Windows 2000 Server. Our DB
> is Oracle 9i running on another server. We've had problems quite a few
> times on the database server where it either has rebooted or been taken
> down for maintenance. In these instances I have to completely restart
> the Remedy server. I can't even just stop and start the service.
>
> Is this normal or should the Remedy service reconnect to the database
> once it is back up?
>
> Thanks,
> James
>
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org


20060125This posting was submitted with HTML in it

20060125This posting was submitted with HTML in it


20060125This posting was submitted with HTML in it

20060125This posting was submitted with HTML in it

**

There is already an ar.cfg tag called 'Db-Connection-Retries' that will cause AR Server to attempt to reconnect to the database the specified number of times, in 15 second intervals. Although it hasn't always worked for us.



For example, to have it try to reconnect every 15 seconds over a 20 minute period (80 attempts) you would use:



Db-Connection-Retries: 80



Again, this hasn't worked well for us. It worked initially in our dev server but never worked in our prod system. It seems like our production servergroup thread, email engine thread, or some other thread will always terminate AR Server before the retries are finished.



Chad Hall
(501) 342-2650



From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG] On Behalf Of McKenzie, James J C-E LCMC HQISEC/L3
Sent: Tuesday, March 14, 2006 7:59 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?



Michiel:

I would not like this feature. Why? Because there is a retry limit in the ar.conf file. This forces the arserverd process on Unix and arserver.exe programs to terminate whenever the database fails. This is a safety feature that prevents certain actions from happening. However, this does not always happen and if users are logged in when the database fails can cause miscommunications. What I would like to see is a fix for a long standing problem and that is the error message generation process. If you cannot communicate with the database, then the error message should state so like this:

"The Remedy Action Request Server cannot communicate with its associated database. Please advise the Remedy Administrator of this difficulty" rather than the cryptic ARERR 91/92/93/94 stuff. This would provide more information than I need in order to fix the problem. This information and what the user was doing at the time will suffice.

If communications is lost with the database, a constantly bouncing ARS process/program would be more frustrating to my users than a dead system.

James McKenzie

-----Original Message-----
From: Action Request System discussion list(ARSList) [mailto:arslist@ARSLIST.ORG]On Behalf Of Michiel Beijen
Sent: Tuesday, March 14, 2006 6:34 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?



** Of course it would be far better that the AR System service would try to reconnect to the database every x minutes when it can not connect to the database. But I guess this would be an enhancement request.

Shal we all submit this enhancement request?

Regards,
Michiel



On 3/14/06, Krishnakumar wrote:
Hi James,
Like Brian has mentioned, the ideal thing to do would be to
monitor the DB and stop the ARS service if the DB is down.

However, if this turns out to be a bit out of reach, you could atleast
make sure that ARS is stopped before the DB is brought up again. This
would ensure that ARS is not using the old connection to the DB. ARS can
then be started once the SQL Listeners are up and running. This should
ensure that ARS works fine.



Regards,
Krishnakumar

On Tue, 14 Mar 2006 05:44:21 +0530, Luke, Brian < Brian.Luke@SCHWAB.COM>
wrote:

> James,
> We ran in to this situation as well. Our specific config was
> running Windows app server and Linux DB server.
>
> The best thing to do is to stop ARS on the remote machine (from the DB
> machine) as part of the script stopping the database. Likewise starting
> ARS just after starting the DB. If they're on the same platform it's a
> bit simpler. If it's not on the same platform, you'll need to figure
> out how to issue commands remotely across platforms.
>
> What platforms are your respective systems on?
> Do you have Tivoli, NetIQ, or any other monitoring systems in place the
> systems?
>
> In our case, we installed a cygwin/SSH server on the Windows app server,
> and can issue commands from the Linux box that way.
>
> Another approach that works is to create a monitor to test the return
> code from TNSPING on the app server to your DB Instance. When it fails,
> stop ARS. When it works (and didn't last time it checked), start ARS.
>
> Hope this helps,
> -Brian
>
> -----Original Message-----
> From: Action Request System discussion list(ARSList)
> [mailto:arslist@ARSLIST.ORG] On Behalf Of James Pifer
> Sent: Monday, March 13, 2006 1:43 PM
> To: arslist@ARSLIST.ORG
> Subject: Should Remedy reconnect to the database if the connection is
> lost and then restored?
>
> We're running ARS 6.x with Helpdesk 4.0.3 on Windows 2000 Server. Our DB
> is Oracle 9i running on another server. We've had problems quite a few
> times on the database server where it either has rebooted or been taken
> down for maintenance. In these instances I have to completely restart
> the Remedy server. I can't even just stop and start the service.
>
> Is this normal or should the Remedy service reconnect to the database
> once it is back up?
>
> Thanks,
> James
>
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org



20060125This posting was submitted with HTML in it

*************************************************************************
The information contained in this communication is confidential, is
intended only for the use of the recipient named above, and may be
legally privileged.

If the reader of this message is not the intended recipient, you are
hereby notified that any dissemination, distribution or copying of this
communication is strictly prohibited.

If you have received this communication in error, please resend this
communication to the sender and delete the original message or any copy
of it from your computer system.

Thank you.
*************************************************************************
20060125This posting was submitted with HTML in it 20060125This posting was submitted with HTML in it

I guess the answer depends on who you ask. If you ask me, Remedy
should automatically reconnect to the database when it comes back
up, but in reality it does not. It used to do so a long time ago...
perhaps in version 3. I complained to Remedy tech support when
this feature was taken away, and they said that a customer requested
that Remedy not try to automatically reconnect. I don't understand
that customer's logic, but they apparently have more influence than I do.

In our environment, we have Remedy and Oracle running on separate
boxes. I requested new equipment back in 2003, with the assumption
that both boxes would continue to remain on the same subnet. We had a
few network hiccups shortly before I ordered the equipment, so I went
ahead and ordered two GigE cards to have a private network segment
between both of the new boxes. Shortly after receiving the equipment,
I found out that all of the database servers would be moved to a
different subnet, and each subnet would be firewalled. If it wasn't
for the private network segment, I think we would have had a lot of
Remedy stability issues. Luckily, everything has run smoothly using
a private network segment with maybe one hiccup every year or two.

When we take Oracle down for a backup, a script sshes into the
Remedy box to tell ARSystem to stop. After the backup completes,
the script restarts Remedy.

In your situation, it sounds like the root cause of your problem is
people related as opposed to being hardware or network related.
As another poster mentioned, see if you can somehow get the database
server to shut down Remedy before your database goes down for maintenance.
The issue with you having to reboot your entire box instead of just
stopping and restarting the Remedy service sounds like overkill. I'll
refrain from making any Windows vs Unix comments.

Alex

On Mon, Mar 13, 2006 at 04:43:16PM -0500, James Pifer wrote:
> We're running ARS 6.x with Helpdesk 4.0.3 on Windows 2000 Server. Our DB
> is Oracle 9i running on another server. We've had problems quite a few
> times on the database server where it either has rebooted or been taken
> down for maintenance. In these instances I have to completely restart
> the Remedy server. I can't even just stop and start the service.
>
> Is this normal or should the Remedy service reconnect to the database
> once it is back up?
>
> Thanks,
> James
>
>
> UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org

--
O--------------O--O---------------------------------O
| Alex Agle \/ Systems Support Specialist III |
O---------------/\ Georgia Institute of Technology |
|(404)894-6165 //\\ ITS-Collaborative Applications |
| Atlanta, GA //\/\\ alex.agle(@)oit.gatech.edu |
O------------O-\/\/-O-------------------------------O


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org

On Tue, 2006-03-14 at 17:59 +0530, Krishnakumar wrote:
> Hi James,
> Like Brian has mentioned, the ideal thing to do would be to
> monitor the DB and stop the ARS service if the DB is down.
>
> However, if this turns out to be a bit out of reach, you could atleast
> make sure that ARS is stopped before the DB is brought up again. This
> would ensure that ARS is not using the old connection to the DB. ARS can
> then be started once the SQL Listeners are up and running. This should
> ensure that ARS works fine.
>
>
> Regards,
> Krishnakumar

>From all the posts here's what I've gathered.

Obviously for a scheduled Oracle downtime it's best to stop Remedy
before the DBAs take down Oracle. This is easily accomplished.

For those times when the database is down unexpectedly, or unreachable,
Remedy needs to be restarted, but there's question to whether you can
just restart the service or not.

Sounds like Remedy possibly tries to use an old connection until you
restart? That would back the experience I have had. Restarting the
service does not fix our problem.

For the record, our Remedy server is:
Windows 2000 SP4, current updates
ARS 6.03.00 Patch 11
Helpdesk 4.0.3
Oracle Client 9.2.0.1.0

Oracle Server:
Windows 2000 SP4, current updates
Oracle Server 9.2 (still waiting for the exact version number)

I like the idea of watching the database connection, with tnsping or
whatever, but in our case, since we can't just restart the service it
wouldn't work for us...yet.

In response to the last reply by Alex: Windows vs Unix. Believe me, I
wish we were on Unix or some flavor of supported *nix. Unfortunately we
aren't so I have to deal with what we have. Also, the root cause is the
unexpected problems. For one reason or another the Oracle server has had
some problems in the last six months. It's been down unexpectedly a
handful of times.

Thanks,
James


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org

**

James:

Sadly, this sounds like a Windows problem. I can, and have, HUP'd the arserverd process which caused all of the other processes to fail with a TERM and this brought backup the Remedy server and reconnected it with an Oracle database on another server. Sadly, I think that this might be solved by stopping the Remedy service and waiting a few minutes for all of the communications to die down (I've had to do this with other products) and then restart the service. Have you tried this?

James McKenzie

-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:arslist@ARSLIST.ORG]On Behalf Of James Pifer
Sent: Tuesday, March 14, 2006 8:51 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection
is lost and then restored?


On Tue, 2006-03-14 at 17:59 +0530, Krishnakumar wrote:
> Hi James,
> Like Brian has mentioned, the ideal thing to do would be to
> monitor the DB and stop the ARS service if the DB is down.
>
> However, if this turns out to be a bit out of reach, you could atleast
> make sure that ARS is stopped before the DB is brought up again. This
> would ensure that ARS is not using the old connection to the DB. ARS can
> then be started once the SQL Listeners are up and running. This should
> ensure that ARS works fine.
>
>
> Regards,
> Krishnakumar

>From all the posts here's what I've gathered.

Obviously for a scheduled Oracle downtime it's best to stop Remedy
before the DBAs take down Oracle. This is easily accomplished.

For those times when the database is down unexpectedly, or unreachable,
Remedy needs to be restarted, but there's question to whether you can
just restart the service or not.

Sounds like Remedy possibly tries to use an old connection until you
restart? That would back the experience I have had. Restarting the
service does not fix our problem.

For the record, our Remedy server is:
Windows 2000 SP4, current updates
ARS 6.03.00 Patch 11
Helpdesk 4.0.3
Oracle Client 9.2.0.1.0

Oracle Server:
Windows 2000 SP4, current updates
Oracle Server 9.2 (still waiting for the exact version number)

I like the idea of watching the database connection, with tnsping or
whatever, but in our case, since we can't just restart the service it
wouldn't work for us...yet.

In response to the last reply by Alex: Windows vs Unix. Believe me, I
wish we were on Unix or some flavor of supported *nix. Unfortunately we
aren't so I have to deal with what we have. Also, the root cause is the
unexpected problems. For one reason or another the Oracle server has had
some problems in the last six months. It's been down unexpectedly a
handful of times.

Thanks,
James


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org

20060125This posting was submitted with HTML in it

**
If this is a windows problem, why dont you create a dependency between the ARS Server and the Oracle database wherein the ARS Server service would depend on the Oracle service and would not start without the Oracle services being up.. that way the Oracle database admins wont able to stop the Oracle server without first killing the ARS Server...

These dependencies can be created via the registry - if you do not know how let me know and I can guide you through it...

Regards

Joe D'Souza
Remedy Developer / Consultant,
Shyle Networks,
New Jersey.

"McKenzie, James J C-E LCMC HQISEC/L3" wrote:

**
James:
Sadly, this sounds like a Windows problem. I can, and have, HUP'd the arserverd process which caused all of the other processes to fail with a TERM and this brought backup the Remedy server and reconnected it with an Oracle database on another server. Sadly, I think that this might be solved by stopping the Remedy service and waiting a few minutes for all of the communications to die down (I've had to do this with other products) and then restart the service. Have you tried this?
James McKenzie
-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:arslist@ARSLIST.ORG]On Behalf Of James Pifer
Sent: Tuesday, March 14, 2006 8:51 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection
is lost and then restored?

On Tue, 2006-03-14 at 17:59 +0530, Krishnakumar wrote:
> Hi James,
> Like Brian has mentioned, the ideal thing to do would be to
> monitor the DB and stop the ARS service if the DB is down.
>
> However, if this turns out to be a bit out of reach, you could atleast
> make sure that ARS is stopped before the DB is brought up again. This
> would ensure that ARS is not using the old connection to the DB. ARS can
> then be started once the SQL Listeners are up and running. This should
> ensure that ARS works fine.
>
>
> Regards,
> Krishnakumar
>From all the posts here's what I've gathered.
Obviously for a scheduled Oracle downtime it's best to stop Remedy
before the DBAs take down Oracle. This is easily accomplished.
For those times when the database is down unexpectedly, or unreachable,
Remedy needs to be restarted, but there's question to whether you can
just restart the service or not.
Sounds like Remedy possibly tries to use an old connection until you
restart? That would back the experience I have had. Restarting the
service does not fix our problem.
For the record, our Remedy server is:
Windows 2000 SP4, current updates
ARS 6.03.00 Patch 11
Helpdesk 4.0.3
Oracle Client 9.2.0.1.0
Oracle Server:
Windows 2000 SP4, current updates
Oracle Server 9.2 (still waiting for the exact version number)
I like the idea of watching the database connection, with tnsping or
whatever, but in our case, since we can't just restart the service it
wouldn't work for us...yet.
In response to the last reply by Alex: Windows vs Unix. Believe me, I
wish we were on Unix or some flavor of supported *nix. Unfortunately we
aren't so I have to deal with what we have. Also, the root cause is the
unexpected problems. For one reason or another the Oracle server has had
some problems in the last six months. It's been down unexpectedly a
handful of times.
Thanks,
James



Yahoo! Mail
Use Photomail to share photos without annoying attachments. 20060125This posting was submitted with HTML in it

** This would work obviously - but only if the DB is on the same box as the AR Server....

Regards,
Michiel


On 3/14/06, Joe DeSouza < joeremedy@yahoo.com > wrote:

**
If this is a windows problem, why dont you create a dependency between the ARS Server and the Oracle database wherein the ARS Server service would depend on the Oracle service and would not start without the Oracle services being up.. that way the Oracle database admins wont able to stop the Oracle server without first killing the ARS Server...

These dependencies can be created via the registry - if you do not know how let me know and I can guide you through it...

Regards

Joe D'Souza
Remedy Developer / Consultant,
Shyle Networks,
New Jersey.


"McKenzie, James J C-E LCMC HQISEC/L3" wrote:


**
James:
Sadly, this sounds like a Windows problem. I can, and have, HUP'd the arserverd process which caused all of the other processes to fail with a TERM and this brought backup the Remedy server and reconnected it with an Oracle database on another server. Sadly, I think that this might be solved by stopping the Remedy service and waiting a few minutes for all of the communications to die down (I've had to do this with other products) and then restart the service. Have you tried this?
James McKenzie
-----Original Message-----
From: Action Request System discussion list(ARSList)
[ mailto:arslist@ARSLIST.ORG ]On Behalf Of James Pifer
Sent: Tuesday, March 14, 2006 8:51 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection
is lost and then restored?

On Tue, 2006-03-14 at 17:59 +0530, Krishnakumar wrote:
> Hi James,
> Like Brian has mentioned, the ideal thing to do would be to
> monitor the DB and stop the ARS service if the DB is down.
>
> However, if this turns out to be a bit out of reach, you could atleast
> make sure that ARS is stopped before the DB is brought up again. This
> would ensure that ARS is not using the old connection to the DB. ARS can
> then be started once the SQL Listeners are up and running. This should
> ensure that ARS works fine.
>
>
> Regards,
> Krishnakumar
>From all the posts here's what I've gathered.
Obviously for a scheduled Oracle downtime it's best to stop Remedy
before the DBAs take down Oracle. This is easily accomplished.
For those times when the database is down unexpectedly, or unreachable,
Remedy needs to be restarted, but there's question to whether you can
just restart the service or not.
Sounds like Remedy possibly tries to use an old connection until you
restart? That would back the experience I have had. Restarting the
service does not fix our problem.
For the record, our Remedy server is:
Windows 2000 SP4, current updates
ARS 6.03.00 Patch 11
Helpdesk 4.0.3
Oracle Client 9.2.0.1.0
Oracle Server:
Windows 2000 SP4, current updates
Oracle Server 9.2 (still waiting for the exact version number)
I like the idea of watching the database connection, with tnsping or
whatever, but in our case, since we can't just restart the service it
wouldn't work for us...yet.
In response to the last reply by Alex: Windows vs Unix. Believe me, I
wish we were on Unix or some flavor of supported *nix. Unfortunately we
aren't so I have to deal with what we have. Also, the root cause is the
unexpected problems. For one reason or another the Oracle server has had
some problems in the last six months. It's been down unexpectedly a
handful of times.
Thanks,
James





Yahoo! Mail
Use Photomail to share photos without annoying attachments. 20060125This posting was submitted with HTML in it




20060125This posting was submitted with HTML in it

On Tue, 2006-03-14 at 09:01 -0700, McKenzie, James J C-E LCMC HQISEC/L3
wrote:
> **
>
> James:
>
> Sadly, this sounds like a Windows problem. I can, and have, HUP'd the
> arserverd process which caused all of the other processes to fail with
> a TERM and this brought backup the Remedy server and reconnected it
> with an Oracle database on another server. Sadly, I think that this
> might be solved by stopping the Remedy service and waiting a few
> minutes for all of the communications to die down (I've had to do this
> with other products) and then restart the service. Have you tried
> this?
>
> James McKenzie

No, I have not tried stopping the service and waiting a few minutes and
then restarting. I have a test box that I could try this on and see if
it works. If it does then I could possibly write something that:
watches the connection
if it fails, keep rechecking for it to come back up
when connectivity to Oracle returns:
stop services for some period of time (to be determined)
then restart the services

If I can get it to work in some automated way that would be a workable
solution. If not, restarting the box is just as easy.

In response to Joe DeSouza, Remedy and Oracle are on different servers.

James


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org

Of course - some network latency would explain this. You might use this as an opportunity to have your network guys look at the whole setup, to ensure that settings are as they should be for normal communications as well.


Rick


From: Action Request System discussion list(ARSList) on behalf of James Pifer
Sent: Tue 3/14/2006 8:18 AM
To: arslist@ARSLIST.ORG
Subject: Re: Should Remedy reconnect to the database if the connection is lost and then restored?





On Tue, 2006-03-14 at 09:01 -0700, McKenzie, James J C-E LCMC HQISEC/L3
wrote:
> **
>
> James:
>
> Sadly, this sounds like a Windows problem. I can, and have, HUP'd the
> arserverd process which caused all of the other processes to fail with
> a TERM and this brought backup the Remedy server and reconnected it
> with an Oracle database on another server. Sadly, I think that this
> might be solved by stopping the Remedy service and waiting a few
> minutes for all of the communications to die down (I've had to do this
> with other products) and then restart the service. Have you tried
> this?
>
> James McKenzie

No, I have not tried stopping the service and waiting a few minutes and
then restarting. I have a test box that I could try this on and see if
it works. If it does then I could possibly write something that:
watches the connection
if it fails, keep rechecking for it to come back up
when connectivity to Oracle returns:
stop services for some period of time (to be determined)
then restart the services

If I can get it to work in some automated way that would be a workable
solution. If not, restarting the box is just as easy.

In response to Joe DeSouza, Remedy and Oracle are on different servers.

James


UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org




UNSUBSCRIBE or access ARSlist Archives at http://www.ARSLIST.org