so what did the host actually say during the outage

[grud]

After the three day outage I went through the archived threads from the blood bag incident, at 2am when the host had gone quiet. The pattern is the same. Nobody had authority to make the call because the rules didn't say who was responsible when the server went down mid-session. We lost half the regulars becaues we spent three days arguing about gear preservation instead of just saying the restart was the restart. The blood bag incident at least had a medic and a protocol, even if it failed. This outage was administrative and nobody is laughing. I think we need a single line: admin decides within one hour whether a rollback happens, then everyone moves on. I wrote it up but I'll wait for someone to tell me it's too long to read. It probably is.

[Doc_Mallory]

grud, I read the write-up. The outage was the outage, and no rule would have changed how the host handled those three days. It reminded me of Hollis, honestly — two months of trusting him, then the camp emptied on a Sunday morning. You can't put trust in a line. Still, if adding one line makes the next outage easier for whoever's left, I'm for it. Maybe one hour is generous; the host didn't answer me for two days last time.

edit: wrong year, it was the summer after

[chad]

An hour at 3am jsut means soemone makes the call on no sleep. The tent camp despawn had the same problem from what I heard — nobody could say which admin had the final word. Better to write 'by the next restart' and pin it to whoever's on the roster, even if Doc's right that trust is separate.

[VLK]

grud, I read the whole thing. The one hour is fine but at 3am no one wants to be the one. The ban that split it came up the same way afterward, I only heard about it later and it never really finished. Maybe name the roster admin so it is decided before the host wakes up.

[sniper]

VLK, the roster's only as good as the last time someone updated it. the ban that split it dragged on the same way because nobody could even agree who was admin that week.

[grud]

Sniper has the actual problem: any rule that depends on the roster depends on someone remembering to update the roster, and we have all seen how that goes. The three day outage only went three days because the host's ticket queue was the only authority and nobody could reach it. So make the rule depend on the restart timer instead. If no admin posts a rollback decision before the next scheduled restart, the restart is the rollback and everything is decided. The blood bag incident failed for the same reason: the protocol assumed the bag wouldn't run out and we had no default for when it did. One hour at 3am is still a person making a call at 3am, but a scheduled restart is just a clock. Pin it to that and the roster becomes backup, not the rule. We'll still lose people either way, but at least the ones left won't be re-fighting this in June. If the next outage happens before we add the line, I will point at this thread.

[Doc_Mallory]

Grud, the field taught me the same thing. We had a protocol for the transfusion and none for when it stopped, so we stood in the grsas arguing while pete was out and nobody's plan covered it. A default that doesn't need a person sounds better than a deadline that does, and the scheduled restart is the closest thing we have to a default. I'd keep the roster in the rule as backup, because sniper's right that a list only helps if someone remembers it. The next outage will probably be different anyway, and we'll be back here arguing about that one instead.