I was sticking with a stable version of 19x or 20x for my Mac server, but one of my clients went and updated all their staff computers to the latest FMP client, which doesn't go that far back. I didn't want to update to the latest Filemaker server since it just came out, so I updated to the final version of 21 (21.1.5.500).
I have some overnight scripts that cache a bunch of things and used to take about 15-20 minutes.
Since updating that time has spiked to around 8 hours. Last night it took 18 hours!
It also causes all sorts of weird errors such as "The host's capacity was exceeded try again later" messages, and various operations seemingly getting confused and going sideways. I see 21 had a lot of issues, early on, but it seemed like most of them got better...
There were no other significant changes in that timeframe (you can see my speed tests where I updated the server). Is this a known issue? I'm considering upgrading to the 22, but that feels scary since it JUST released and I don't know what new issues I'll find.
I did go in and put some timed steps in there so I can track down what part(s) are slowing it down the most.
Recursive PSOS scripts? -- in older versions of FMS, calling a PSoS script from within a PSoS script on the server would do nothing. In recent versions, the second script will run. Any chance you have accidental PSoS scripts calling PSoS scripts recursively? This might account for your "host's capacity was exceeded" messages. N.B: This feature was added in FMS 21.1.1, see Claris FileMaker Server Release Notes
Active Directory? In short: one user reports terrible script performance when they have AD connections in FMS 21. See Claris Community (English)
Solved. Wasn't quite that but was the clue I needed to figure it out.
I had a refresh script that would run every time you entered a layout and would trigger itself. It failed gracefully locally, but now breaks on the server.
