Operator guide#

Run Gecko as soon as you notice a problem, before restarting software or power-cycling hardware. It records what the machine looked like at that moment.

File a report#

Open a terminal on the instrument machine and run:

$ gecko -u "your.name" -m "what happened"

Both parts are required:

-u, --user

Your name, so engineers know who to follow up with.

-m, --message

A description of the problem. Put it in quotes.

A run takes a few seconds, most of it spent sampling CPU usage. When it finishes, Gecko prints where the bundle was saved:

Triage workflow complete! Report: /home/obs/gecko_reports/2026-11-06/gecko_2026-11-06_071305_1234560000.tar.gz

If email alerts are configured, the bundle is also sent to the engineering team. Nothing else is required of you.

Writing a useful message#

Engineers read your message long after the event. A good one answers:

  • What did you see? The exact error text, or what looked wrong.

  • When? The UT time, or which exposure or observation it was during.

  • What were you doing? The last command, script or GUI action before the problem.

  • What have you tried? For example, “retried twice, same result”.

$ gecko -u "Steven.B" -m "Slicecam frames all zeros since ~02:10 UT. Started after \
  changing the exposure to 30 s. Retried twice. ACAM looks normal."

If Gecko itself reports an error#

Message

What to do

No configuration file found ... Run: gecko -init

Gecko hasn’t been set up for your account. Ask your instrument administrator, or see installation.

... has not been initialized

Same as above.

you need to include a user and a message

Add both -u and -m.

gecko: command not found

Gecko isn’t on your PATH. Ask your administrator.

Anything else: see troubleshooting. Include the full error text when you report the problem.