Lab briefing · Fill it up on purpose

EXHAUST a tablespace and an archive destination deliberately, observe EXACTLY WHAT USERS SEE in each case, and RECOVER. The two failures look completely different from a user's side and identical from a monitoring dashboard, which is why doing both is worth an afternoon. FILL THE TABLESPACE FIRST, BECAUSE IT IS THE GENTLER ONE. Insert until it cannot extend. Then go and be a user: try to save something in the affected module, and try to work in an unaffected one. The result is the point — one transaction fails, one person sees an error, and everybody else carries on. It is a partial failure, and it escalates slowly, which is why it is often reported hours after it started. THEN FILL THE ARCHIVE DESTINATION, AND EXPECT SOMETHING QUALITATIVELY DIFFERENT. The database stops writing entirely. Sessions do not error — they HANG, mid-transaction, with no message. Go and be a user…

In this module: Module 12 · Space, growth and purging

1. 1What fills up, and in what order
2. 2Monitoring space before it matters
3. 3Growing a tablespace
4. 4The standard purge programs
5. 5Deciding what history to keep
6. 6Purging safely the first time
7. 7Reclaiming space, and when you cannot
8. 8What breaks: three space failures
9. 9Lab briefing · Fill it up on purpose