Any idea? For some reason starting/stopping *any* services  in
SharePoint, hang  on "starting" and/or "stopping' for about 1 hour before
moving to a started/stopped status... so they do eventually get to the
proper state. This problem does not occur on the box that setup the
farm. It happens on every other box though that joins the farm. There
are no errors in the logs or event viewer.

Is there anything in the event logs on the server  or.... is the below
the only event log message - also, do you have the event id number that
is being logged ?

You can also read John Raynors blog
(http://blogs.conchango.com/johnrayner/) where he had this same error  -
he just set it to use the network service  account and they start  up -
you could try that and see if they start and then use the tool to change
the services  again to use the correct account  information in case
something has got corrupted along the way.

Personally, I would probably try and use the service account reset page
first just ion case somethings gone screwy somewhere

It appears the "hanging" service  is being caused by some process trying
to communicate with the database and then eventually times out.
Initially, I did not see anything in event viewer because an error  is
not logged until the timeout occurs - which is just under an hour. Yet
the services  still manage to start  after the timeout...strange.

There are two eventid errors. Note, I have three web front-ends and the
errors are happening  on everyone of them if I try to start the search
services on a particular box. This issue does not occur on the box that
setup the farm  (central admin). I even tried pointing to a different SQL
instance on a different machine and same error occurs.

I am wondering if this is an OS image issue. Details of the eventid
errors are below.

1)

Event Type: Error
Event Source: Windows sharepoint  Services 3
Event Category: Timer
Event ID: 6398
Date: 3/21/2007
Time: 11:41:15 AM
User: N/A
Computer: ZTCFH0JP
Description:
The Execute method of job definition
Microsoft.SharePoint.Administration.SPConfigurationRefreshJobDefinition
(ID 8bc4f35c-87de-4f87-b405-ec26b92db13f) threw an exception. More
information is included below.

A transport-level error has occurred when sending the request to the
server. (provider: TCP Provider, error: 0 - An existing connection was
forcibly closed by the remote host.)

All the stuff i'm reading suggests that its a timeout talking to the sql
server
I take it you've disabled named pipes on the sql server

Are you using Kerberos authentication or ntlm for the portal ? If using
Kerberos have you set the spn (setspn.exe) for the sql database ?

I'm also wondering if you're doing anything that is seizing the thread
in the web app / app pool - you may want to try adding a thread to the
app pool concerned (I take it you've set them up as web gardens to use
at least 2 threads and preferably at least 3) - this might explain why
the wfe ends time out (because of the app pool thing) while central
admin doesn't (as its on a different app pool and its threads aren't
being tied up)

This is a fresh install of OS and SharePoint. After running the product
and technologies wizard to establish a farm, i then login to the second
box and join the farm. Stopping/Starting any service  on the joined
server is taking an hour to work.

I just had a thought. All of our boxes had a DST patch applied recently.
I suppose its likely coincidence that its taking 1 hour to start  but Im
going to speak with a few people.

You didn't answer Philip's question about NTLM vs Kerberos. For Kerberos
you'll need to set up an SPN for the WFE. The SQL Server SPN should be
there automatically for the domain  account running SQL Server.

Well, I think I've been bitten by DST. Apparently, MS is getting a few
calls about this....

http://support.microsoft.com/kb/932563

Testing the resolution now..

that just looks like a normal DST problem tho - as the hour
changes, if its using a tick count to time the jobs, which the timer
service might, the jobs will run one hour after they're scheduled until
the clocks revers when they'll run on time

Let us know how you get on tho - clearly this is something to do with
the update ..... i just dont see why it should take an hour for the
connection to time out unless it keeps replaying for the hour until
clocks match

I take it the DST patch was applied to all servers and that you've
logged on as the local administrator (not a local admin  or domain  admin
but THE local admin) on each box and checked what the clock says ?

I noticed the same thing on our sharepoint  Test Environment but when
I went to start  the service  Via the admin  pages it started just fine.
Just remember to use the domain  name/username format when it asks for
accounts or it will give a cryptic error  message something similar to
(spsearch) Username you tried to run the service with through the
admin pages. Is their a specific reason you need to start the service
manually instead of via the central  administration page?