Detect errors that keep users away??

so other than load balancing and clustering, is there something I can use as
a poor man? what can be that cheap 3rd entity above these two servers?

"David Wang [Msft]" wrote:

 > Not possible. The server that "halts" is by definition unable to detect this
 > state, and thus not possible to fail over.
 >
 > You pretty much need a third entity that monitors the other two servers and
 > performs such a switch, most likely at the DNS level.
 >
 > For example, if server A halts, you don't want people still going to server
 > A and ask it to redirect to server B (what if server A is unable to
 > answer -- your failover just failed to redirect). You want people directly
 > going to server B after server A fails. This is best done with a third
 > entity unrelated to server A and server B, monitoring both servers, and then
 > swapping DNS as appropriate.
 >
 > If you don't want to swap DNS, then you will need a third server C, which is
 > independent from server A and server B, and only tell people about server C.
 > Server C will pass-thru requests to server A, monitor its uptime, and if
 > down, transparently switch to server B.
 >
 > But, if you had server C, you are no longer using a "poor man's
 > fail-over" -- that's what real failover solutions do.
 >
 > --
 > //David
 > IIS
<font color=purple> > <a style='text-decoration: underline;' href="http://blogs.msdn.com/David.Wang</font" target="_blank">http://blogs.msdn.com/David.Wang</font</a>>
 > This posting is provided "AS IS" with no warranties, and confers no rights.
 > //
 > "rweide" <rweide.DeleteThis@discussions.microsoft.com> wrote in message
 > news:A677E2B6-0CEB-4596-89E6-357D0235CD98@microsoft.com...
 > I use one of two small servers to host my intranet website, the other as a
 > backup.
 >
 > I am trying to detect server errors (any error that halts the server) in my
 > classic ASP code, so that when error occurs in one server, I will redirect
 > the process to the other server - sort of a poor man's web farm to failover.
 >
 > How do I detect this type of errors that stop the users from accessing the
 > web site?
 >
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Detect errors that keep users away?? 

Not really. A "cheap 3rd entity above these two servers" is essentially the
load balancer, so if you do not want to purchase the necessary equipment,
you will have to buy/find the right software, a new PC to run it on,
configure the software properly, and maintain an extra machine.

Thus, as soon as you require resources for a new PC in the game, it is
cheaper to just get a load balancer that is well suited for the task.

Personally, as soon as you are worrying about "availability", it tells me
that you should be handing the whole thing over to centralized IT where
people are paid to keep things available. You do not need to do their job
(unless you are considered "IT"). You are asking for additional hardware and
specialized software, and there is no "poor man's" way about it.

In either case, if there is demand for you to keep the server "available",
explain that funds are necessary to accomplish that. You don't get something
for nothing, so if there aren't any funds and you do not have the time to
configure/maintain a third machine, just let it drop -- people may scream,
but it just becomes ammo for you to get funds for the right hardware to
satisfy the users' requirements.

--
//David
IIS
<a style='text-decoration: underline;' href="http://blogs.msdn.com/David.Wang" target="_blank">http://blogs.msdn.com/David.Wang</a>
This posting is provided "AS IS" with no warranties, and confers no rights.
//
"rweide" <rweide.TakeThisOut@discussions.microsoft.com> wrote in message
news:B4CD18EA-3992-4338-A395-4FAB87590A32@microsoft.com...
so other than load balancing and clustering, is there something I can use as
a poor man? what can be that cheap 3rd entity above these two servers?

"David Wang [Msft]" wrote:

 > Not possible. The server that "halts" is by definition unable to detect
this
 > state, and thus not possible to fail over.
 >
 > You pretty much need a third entity that monitors the other two servers
and
 > performs such a switch, most likely at the DNS level.
 >
 > For example, if server A halts, you don't want people still going to
server
 > A and ask it to redirect to server B (what if server A is unable to
 > answer -- your failover just failed to redirect). You want people directly
 > going to server B after server A fails. This is best done with a third
 > entity unrelated to server A and server B, monitoring both servers, and
then
 > swapping DNS as appropriate.
 >
 > If you don't want to swap DNS, then you will need a third server C, which
is
 > independent from server A and server B, and only tell people about server
C.
 > Server C will pass-thru requests to server A, monitor its uptime, and if
 > down, transparently switch to server B.
 >
 > But, if you had server C, you are no longer using a "poor man's
 > fail-over" -- that's what real failover solutions do.
 >
 > --
 > //David
 > IIS
<font color=purple> > <a style='text-decoration: underline;' href="http://blogs.msdn.com/David.Wang</font" target="_blank">http://blogs.msdn.com/David.Wang</font</a>>
 > This posting is provided "AS IS" with no warranties, and confers no
rights.
 > //
 > "rweide" <rweide.TakeThisOut@discussions.microsoft.com> wrote in message
 > news:A677E2B6-0CEB-4596-89E6-357D0235CD98@microsoft.com...
 > I use one of two small servers to host my intranet website, the other as a
 > backup.
 >
 > I am trying to detect server errors (any error that halts the server) in
my
 > classic ASP code, so that when error occurs in one server, I will redirect
 > the process to the other server - sort of a poor man's web farm to
failover.
 >
 > How do I detect this type of errors that stop the users from accessing the
 > web site?
 >
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Detect errors that keep users away?? 

good information and suggestion, every words understood and well received. =)

Thanks.

"David Wang [Msft]" wrote:

 > Not really. A "cheap 3rd entity above these two servers" is essentially the
 > load balancer, so if you do not want to purchase the necessary equipment,
 > you will have to buy/find the right software, a new PC to run it on,
 > configure the software properly, and maintain an extra machine.
 >
 > Thus, as soon as you require resources for a new PC in the game, it is
 > cheaper to just get a load balancer that is well suited for the task.
 >
 > Personally, as soon as you are worrying about "availability", it tells me
 > that you should be handing the whole thing over to centralized IT where
 > people are paid to keep things available. You do not need to do their job
 > (unless you are considered "IT"). You are asking for additional hardware and
 > specialized software, and there is no "poor man's" way about it.
 >
 > In either case, if there is demand for you to keep the server "available",
 > explain that funds are necessary to accomplish that. You don't get something
 > for nothing, so if there aren't any funds and you do not have the time to
 > configure/maintain a third machine, just let it drop -- people may scream,
 > but it just becomes ammo for you to get funds for the right hardware to
 > satisfy the users' requirements.
 >
 > --
 > //David
 > IIS
<font color=purple> > <a style='text-decoration: underline;' href="http://blogs.msdn.com/David.Wang</font" target="_blank">http://blogs.msdn.com/David.Wang</font</a>>
 > This posting is provided "AS IS" with no warranties, and confers no rights.
 > //
 > "rweide" <rweide.RemoveThis@discussions.microsoft.com> wrote in message
 > news:B4CD18EA-3992-4338-A395-4FAB87590A32@microsoft.com...
 > so other than load balancing and clustering, is there something I can use as
 > a poor man? what can be that cheap 3rd entity above these two servers?
 >
 > "David Wang [Msft]" wrote:
 >
  > > Not possible. The server that "halts" is by definition unable to detect
 > this
  > > state, and thus not possible to fail over.
  > >
  > > You pretty much need a third entity that monitors the other two servers
 > and
  > > performs such a switch, most likely at the DNS level.
  > >
  > > For example, if server A halts, you don't want people still going to
 > server
  > > A and ask it to redirect to server B (what if server A is unable to
  > > answer -- your failover just failed to redirect). You want people directly
  > > going to server B after server A fails. This is best done with a third
  > > entity unrelated to server A and server B, monitoring both servers, and
 > then
  > > swapping DNS as appropriate.
  > >
  > > If you don't want to swap DNS, then you will need a third server C, which
 > is
  > > independent from server A and server B, and only tell people about server
 > C.
  > > Server C will pass-thru requests to server A, monitor its uptime, and if
  > > down, transparently switch to server B.
  > >
  > > But, if you had server C, you are no longer using a "poor man's
  > > fail-over" -- that's what real failover solutions do.
  > >
  > > --
  > > //David
  > > IIS
<font color=green>  > > <a style='text-decoration: underline;' href="http://blogs.msdn.com/David.Wang</font" target="_blank">http://blogs.msdn.com/David.Wang</font</a>>
  > > This posting is provided "AS IS" with no warranties, and confers no
 > rights.
  > > //
  > > "rweide" <rweide.RemoveThis@discussions.microsoft.com> wrote in message
  > > news:A677E2B6-0CEB-4596-89E6-357D0235CD98@microsoft.com...
  > > I use one of two small servers to host my intranet website, the other as a
  > > backup.
  > >
  > > I am trying to detect server errors (any error that halts the server) in
 > my
  > > classic ASP code, so that when error occurs in one server, I will redirect
  > > the process to the other server - sort of a poor man's web farm to
 > failover.
  > >
  > > How do I detect this type of errors that stop the users from accessing the
  > > web site?
  > >
  > >
  > >
 >
 >
 ><!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: Detect errors that keep users away??