Operation Modes & Logon Load Balancing

An instance restarts at 02:00 during the night batch mode. Right after startup the work-process mix matches the profile, not the night mode. Why, and does it self-correct?

Your application servers each have two network cards — a fast link to the database and the general network to the frontends. What must you configure so logons reach the right address?

You need to do service work on one application server without abruptly kicking users off. How do you drain new logons from it transparently?

Parallel RFC processing is overloading one dialog instance. Which mechanism balances that load, and why isn't it SMLG logon groups?

You suspect RFCs aren't running because of resource limits. Which transaction diagnoses it, and what does the Max. Resources figure actually represent?

When a user logs on with load balancing, how does SAP GUI end up connected to the least-loaded application server, given that sapmsg.ini only lists the message server?

You try to start an instance from the Control Panel and it fails, even though the OS-level remote start works. What SAP-side setting is usually missing?

You're defining a new instance's work-process distribution in RZ04 and the numbers look blank or arbitrary. Which step populates the correct starting point?

To build an aggressive night batch mode you drop background processes to 0 but keep the instance's class A background processes. Why can no jobs run there?

Which transaction is the Operation Mode Calendar, and how does it differ from RZ04 and RZ03?

A one-off month-end invoicing run needs a different work-process mix for a single evening. Do you create a new operation mode just for it?

A scheduled operation mode switch fires, but a work process you expected to become a background process is still dialog moments later. Bug?

After defining operation modes you assign them in the RZ04 time table but the switch never happens during some hours. What's the likely mistake, and what grid detail can you set?

On a new system only the DUMMY operation mode exists. Can you build your day/night automatic switching on top of it?

During a day-to-night operation mode switch, an admin expects the total work-process count to grow to handle batch. Why is that the wrong mental model, and what does it mean for capacity planning?

To maximize night-time batch you draft an operation mode that takes dialog processes to 0 and removes the enqueue process. Which of these will the system refuse, and why does it matter?

On a freshly built instance you open RZ04 to define operation modes but the resource distribution looks empty or wrong. What precondition was skipped?

You enable rdisp/dynamic_wp_check=TRUE expecting it to permanently boost process capacity. What does it actually do, and what's the catch about lifetime?

Your overnight batch jobs need more background processes than the daytime config has. A colleague says you must schedule a restart to apply the night-time setup. Why is that wrong?

From RZ03 you can start/stop the local instance, but starting a remote UNIX instance fails. On UNIX, what host-level prerequisite is the first thing to check?

A runtime bottleneck (e.g. an RFC cascade deadlock) is choking the instance because every process is occupied. Why won't an operation mode switch resolve it, and what mechanism actually can?

You set a logon group's 'max average response time' and 'max users' to cap an app server, but users keep logging on past the limit. Bug, or expected?

An incident needs more batch processes right now, ahead of the scheduled night switch. Which transaction lets you force the mode change off-schedule, and how does it differ from where you defined the modes?

A user complains they always land on the same overloaded app server. With logon load balancing via logon groups, is that how assignment actually works?

In SMLG, your team worries the default 'Best' strategy will pile every new logon onto one top server. How does Best actually spread load, and when would you pick Round Robin instead?

A bottleneck hits at runtime and you want to bring the reserved work processes (rdisp/wp_no_restricted) into play. What constraint determines whether that's even possible?

You're sizing a large instance and keep raising rdisp/wp_max_no to fit more work processes. The sum of your process parameters is climbing past 600. What stops you, and what does that ceiling mean?

You raise an RFC server group quota in RZ12, it takes effect immediately, but after the next instance restart the RFCs fail again. What was missed?

You have exactly two application servers and want high availability via a logon group. Is putting both instances in one group on those two servers the right move?