Starting, Stopping & Diagnostics

On a scale-out (distributed) HANA system you need to stop everything. Why is HDB stop the wrong tool, and what do you use instead?

An instance crashed twice; you go to read the dev traces from the first crash and they're gone. Why?

HANA trace files are filling the disk. What two parameters cap them, where do they live, and why is hand-deleting the files the wrong fix?

GetProcessList shows GREEN across the board, but users report the dialog instance is hung. What is the trap, and what does YELLOW actually warn you about?

An RFC/gateway problem is suspected. Among dev_disp, dev_w#, dev_rd, dev_icm, dev_ms, which file do you open, and which name is the easy mix-up?

An admin connects to a tenant database to start/stop the whole HANA system and the command fails. What is the contraindication?

You must triage a noisy HANA trace fast. Among the severity letters d/i/w/e, which do you grep for first, and once you find it how do you tie it back to the offending session?

To capture a transient issue you're tempted to set the dev trace to its highest level and leave it. Why is that the wrong default, and what level is safe to run continuously?

For HANA specifically, what does a soft shutdown wait for, and how does its default timeout differ from the application-server soft shutdown?

You issue a soft shutdown from the SAP Management Console during a maintenance window that ends in 20 minutes. Why might users still get cut off, and what decides it?

A colleague says 'reset the dev trace, we can still read the old data afterward.' Why are they wrong?

You need the developer trace specifically for the message server (not the dispatcher). Which transaction, and what is the discriminator from the work-process trace?

During a full system stop, why are dialog instances stopped first and the ASCS last — and what breaks if you reverse it?

Your sapcontrol command fails with permission/connection issues. As which user must it run, and what does that user actually talk to?

The instance is running but the SAP GUI is completely unresponsive — every work process is occupied. Before you reach for a restart, what command-line tool gets you in?

You're hunting a specific tenant's HANA trace. Which directory, and what separates tenant traces from system traces?

sapcontrol returns FAIL: NIECONN_REFUSED. What is the root cause, and what do you check before retrying the function?

You want to automate SAP start/stop on Windows but think the MMC console is the only way. What's the misconception?

Via hdbsql you try ALTER SYSTEM STOP DATABASE for a tenant and get a privilege error. Where must you run it, and which privilege is actually required?

HANA had an incident and you have minutes to find the cause. Why open the alert trace before the full service trace, and what's the catch?

HANA failed to recover and you need the right dump for support. Which file is the crash artifact, and which dump types are easy to grab by mistake?

You want to raise the trace level for the gateway alone, not every facility. Why is RZ11 / rdisp/TRACE the wrong lever, and what is the right one?

Right after a server reboot, StartSystem fails. What precondition is missing, and why does stopping an instance NOT stop the related service?

An old runbook tells you to run startsap/stopsap on an S/4HANA host. Why will that fail, and what replaced it?

A dev trace reaches a size limit during an incident. With rdisp/TRACE_LOGGING at its default, do you lose the new entries or the old ones?

You need an ABAP instance's dev traces for one specific work process. Which directory holds them, and how do you know which dev_ file is that process?

Cleaning up the HANA trace directory to reclaim space, which two files must you never delete, and why?

SAP support asks for diagnostics spanning a specific incident window. Rather than tarring up everything, what does HANA Studio do for you?

You schedule a tenant database stop and assume a soft option will let in-flight work finish. What will actually happen regardless of settings?

HANA has stopped answering SQL, so DBA Cockpit is useless because it runs on the affected AS ABAP. How do you get back in and clear the blockage?

After a system update, your custom timeout tweaks to the SAP systemd service files vanished. What did you do wrong, and where should those changes live?

Before shutting S/4HANA down, you want to warn users off, block new logons and jobs, and cut Fiori access — without locking yourself out. Which transaction do you use, and what safeguard do you need?

You need to monitor HANA but the graphical tools are unavailable. If the S/4HANA system is up, what's the next-best option?

On a Linux HANA host, what actually launches the database at OS boot, and which single profile setting decides whether HANA comes up automatically?

You upgrade an S/4HANA ABAP host to SUSE SLES 15 and notice the instances no longer come up the old sapinit way at boot. What start/stop framework is now in control, and how do you confirm it is active?

On a systemd-managed Linux host you technically can stop an SAP instance with systemctl stop SAP<SID>_<nr>.service. Why does SAP advise against doing so, and what should you use instead?