Forum Search:
Forum.Brain-Cluster.com: Brain Cluster Technical Forum
Ultimate forum for Technical Discussions

Home » Microsoft » Windows Server » Server Clustering » NLB and application failure
NLB and application failure [message #158026] Thu, 30 July 2009 05:20 Go to next message
Mubashir Khan  is currently offline Mubashir Khan
Messages: 6
Registered: July 2009
Junior Member
Dear All,

We have assessed that NLB provides clustering on the machine basis and fail
over takes effect. But what if the machine is ok but our application fails
for some reason.

Is there any way that the incoming requests could be redirected to the nodes
that are running the application.

Thanks
Re: NLB and application failure [message #158028 is a reply to message #158026] Thu, 30 July 2009 09:58 Go to previous messageGo to next message
Russ Kaufmann  is currently offline Russ Kaufmann
Messages: 131
Registered: July 2009
Senior Member
"Mubashir Khan" <MubashirKhan@discussions.microsoft.com> wrote in message
news:AA612A27-7908-4E7D-ACCC-2DE131661CA5@microsoft.com...
> Dear All,
>
> We have assessed that NLB provides clustering on the machine basis and
> fail
> over takes effect. But what if the machine is ok but our application fails
> for some reason.
>
> Is there any way that the incoming requests could be redirected to the
> nodes
> that are running the application.

NLB does not have any application health checking capabilities. If you need
automated failover based on application/service health, you should look at
failover clustering. Of course, failover clustering resource checking
requires cluster aware resources.

--
Russ Kaufmann,
MVP, MCSE: Messaging and Security, MCT, MCITP, MCTS and other stuff

ClusterHelp.com, a Microsoft Certified Gold Partner
Web http://www.clusterhelp.com
Blog http://msmvps.com/clusterhelp
Re: NLB and application failure [message #158050 is a reply to message #158028] Sat, 01 August 2009 18:11 Go to previous messageGo to next message
Garry Starck-MCITP En  is currently offline Garry Starck-MCITP En
Messages: 69
Registered: July 2009
Member
Hi Guys

Hi

Run them in Multicast mode. Write a scipt, Each Node pings the defailt
Gateway every 30 seconds, and establishes a session to say port 80 or 8080
for a nlb proxy, or 3389 for Terminal Services. If the cannot telnet to
themselves on the services port. Then execute a restart of it's own service,
if still not luck, execute the WLBS -??? Suspend commandline to stop itself
answering. I found the script about 3 years back on the net. Some contructive
developer obviously was where your'e at now. However, sometimes the service
still runs in a hund status ans still allows the port to answer every now and
then and this just goes to prove a proper cluster maybe a better option, but
they also seems to Clust F%%^Y sometimes. Even had a send mail alert within
his script.

Have you tried setting the services with restart dependencies / retries in
the properties.

Garry Starck
MCITP Enterprise Administrator, MCTS AD, MCSE 2003 Messaging, MCDBA


"Russ Kaufmann" wrote:

> "Mubashir Khan" <MubashirKhan@discussions.microsoft.com> wrote in message
> news:AA612A27-7908-4E7D-ACCC-2DE131661CA5@microsoft.com...
> > Dear All,
> >
> > We have assessed that NLB provides clustering on the machine basis and
> > fail
> > over takes effect. But what if the machine is ok but our application fails
> > for some reason.
> >
> > Is there any way that the incoming requests could be redirected to the
> > nodes
> > that are running the application.
>
> NLB does not have any application health checking capabilities. If you need
> automated failover based on application/service health, you should look at
> failover clustering. Of course, failover clustering resource checking
> requires cluster aware resources.
>
> --
> Russ Kaufmann,
> MVP, MCSE: Messaging and Security, MCT, MCITP, MCTS and other stuff
>
> ClusterHelp.com, a Microsoft Certified Gold Partner
> Web http://www.clusterhelp.com
> Blog http://msmvps.com/clusterhelp
>
Re: NLB and application failure [message #158051 is a reply to message #158050] Sun, 02 August 2009 02:50 Go to previous messageGo to next message
Vladimir Petter  is currently offline Vladimir Petter  United States
Messages: 15
Registered: August 2009
Junior Member
Also consider using NLB's MOM pack + MOM pack for your application. See
http://blogs.msdn.com/clustering/archive/2009/04/28/9572859. aspx for
details.
Vladimir.
Clustering & HA
Microsoft

"Garry Starck-MCITP Enterprise Admin"
<vjsparx@REMOVE_CAPS_INVALIDhotmail.com> wrote in message
news:BAEF0FA2-8B5E-41D3-8C47-3AE67D91BBB5@microsoft.com...
> Hi Guys
>
> Hi
>
> Run them in Multicast mode. Write a scipt, Each Node pings the defailt
> Gateway every 30 seconds, and establishes a session to say port 80 or 8080
> for a nlb proxy, or 3389 for Terminal Services. If the cannot telnet to
> themselves on the services port. Then execute a restart of it's own
> service,
> if still not luck, execute the WLBS -??? Suspend commandline to stop
> itself
> answering. I found the script about 3 years back on the net. Some
> contructive
> developer obviously was where your'e at now. However, sometimes the
> service
> still runs in a hund status ans still allows the port to answer every now
> and
> then and this just goes to prove a proper cluster maybe a better option,
> but
> they also seems to Clust F%%^Y sometimes. Even had a send mail alert
> within
> his script.
>
> Have you tried setting the services with restart dependencies / retries in
> the properties.
>
> Garry Starck
> MCITP Enterprise Administrator, MCTS AD, MCSE 2003 Messaging, MCDBA
>
>
> "Russ Kaufmann" wrote:
>
>> "Mubashir Khan" <MubashirKhan@discussions.microsoft.com> wrote in message
>> news:AA612A27-7908-4E7D-ACCC-2DE131661CA5@microsoft.com...
>> > Dear All,
>> >
>> > We have assessed that NLB provides clustering on the machine basis and
>> > fail
>> > over takes effect. But what if the machine is ok but our application
>> > fails
>> > for some reason.
>> >
>> > Is there any way that the incoming requests could be redirected to the
>> > nodes
>> > that are running the application.
>>
>> NLB does not have any application health checking capabilities. If you
>> need
>> automated failover based on application/service health, you should look
>> at
>> failover clustering. Of course, failover clustering resource checking
>> requires cluster aware resources.
>>
>> --
>> Russ Kaufmann,
>> MVP, MCSE: Messaging and Security, MCT, MCITP, MCTS and other stuff
>>
>> ClusterHelp.com, a Microsoft Certified Gold Partner
>> Web http://www.clusterhelp.com
>> Blog http://msmvps.com/clusterhelp
>>
Re: NLB and application failure [message #158053 is a reply to message #158051] Sun, 02 August 2009 04:15 Go to previous message
Garry Starck-MCITP En  is currently offline Garry Starck-MCITP En
Messages: 69
Registered: July 2009
Member
Thanks for that tip, I have implemented all AD/SYSVOL MP's, and never even
gave MOM a thought - now I feel stupid
--
Garry Starck
MCITP Enterprise Administrator, MCTS AD, MCSE 2003 Messaging, MCDBA


"Vladimir Petter" wrote:

> Also consider using NLB's MOM pack + MOM pack for your application. See
> http://blogs.msdn.com/clustering/archive/2009/04/28/9572859. aspx for
> details.
> Vladimir.
> Clustering & HA
> Microsoft
>
> "Garry Starck-MCITP Enterprise Admin"
> <vjsparx@REMOVE_CAPS_INVALIDhotmail.com> wrote in message
> news:BAEF0FA2-8B5E-41D3-8C47-3AE67D91BBB5@microsoft.com...
> > Hi Guys
> >
> > Hi
> >
> > Run them in Multicast mode. Write a scipt, Each Node pings the defailt
> > Gateway every 30 seconds, and establishes a session to say port 80 or 8080
> > for a nlb proxy, or 3389 for Terminal Services. If the cannot telnet to
> > themselves on the services port. Then execute a restart of it's own
> > service,
> > if still not luck, execute the WLBS -??? Suspend commandline to stop
> > itself
> > answering. I found the script about 3 years back on the net. Some
> > contructive
> > developer obviously was where your'e at now. However, sometimes the
> > service
> > still runs in a hund status ans still allows the port to answer every now
> > and
> > then and this just goes to prove a proper cluster maybe a better option,
> > but
> > they also seems to Clust F%%^Y sometimes. Even had a send mail alert
> > within
> > his script.
> >
> > Have you tried setting the services with restart dependencies / retries in
> > the properties.
> >
> > Garry Starck
> > MCITP Enterprise Administrator, MCTS AD, MCSE 2003 Messaging, MCDBA
> >
> >
> > "Russ Kaufmann" wrote:
> >
> >> "Mubashir Khan" <MubashirKhan@discussions.microsoft.com> wrote in message
> >> news:AA612A27-7908-4E7D-ACCC-2DE131661CA5@microsoft.com...
> >> > Dear All,
> >> >
> >> > We have assessed that NLB provides clustering on the machine basis and
> >> > fail
> >> > over takes effect. But what if the machine is ok but our application
> >> > fails
> >> > for some reason.
> >> >
> >> > Is there any way that the incoming requests could be redirected to the
> >> > nodes
> >> > that are running the application.
> >>
> >> NLB does not have any application health checking capabilities. If you
> >> need
> >> automated failover based on application/service health, you should look
> >> at
> >> failover clustering. Of course, failover clustering resource checking
> >> requires cluster aware resources.
> >>
> >> --
> >> Russ Kaufmann,
> >> MVP, MCSE: Messaging and Security, MCT, MCITP, MCTS and other stuff
> >>
> >> ClusterHelp.com, a Microsoft Certified Gold Partner
> >> Web http://www.clusterhelp.com
> >> Blog http://msmvps.com/clusterhelp
> >>
>
Goto Forum:
  


Current Time: Thu Sep 21 08:06:35 EDT 2017

Total time taken to generate the page: 0.02540 seconds
.:: Contact :: Home ::Sitemap::.

Powered by: FUDforum 3.0.0RC2.
Copyright ©2001-2009 FUDforum Bulletin Board Software