Time associated with Restore Upon Death in console is 350 seconds. I have three plugins that have not been updated that are above 100 seconds (the other two are above 600 seconds). Really needs an update.
Troubles since Facepunch update 8/3
The aggregate hook time shown by oxide when running o.plugins is the cumulative hook time used since the plugin was loaded (most commonly that's equivalent to when the server started). Sharing that the hook time is 350 seconds without sharing how long the plugin has been loaded, doesn't necessarily indicate that there is a problem. Comparing the cumulative hook time of one plugin with that of other plugins also isn't very useful since I don't know what plugins you are referring to or how often their hooks would be invokved on your servers.
If it's concluded that this plugin has performance issues, we would first have to investigate why (ideally we should be able to reproduce the issue to troubleshoot it easily), before any update can be made to address it. Most commonly, plugins have high cumulative hook time because they subscribe to hooks that are called very frequently, and/or the plugin runs expensive functions when those hooks are called. Glancing at the code of Restore Upon Death, I don't see either taking place.
We do a daily restart at 6am CST. We have periods where our server fps plunges from 60 fps to 7fps. We run plugins command, and then manually reload all those above 100 seconds. We've replaced several plugins with alternatives and obviously updated those with updates, but we still have to reload RestoreUponDeath each time. When we do the o.reload, and we're down to o.reload, truepve, and skilltree as the ones with inordinartely long times (even though skilltree just updated), it then rights itself and goes back to normal... for a while. Then it goes down again, and we repeat. Since it rebounds when we force the reloading of those plugins, I can't help but associate the problem as being those plugins that remain with long associated times. Changing from stacksizecontroller to stack modifier, for instance, went from stacksize being 690 seconds to stackmodifier being 4 seconds.
Just in case there is any confusion, if a plugin shows X seconds of hook time, that doesn't mean it's currently causing lag, that's just a counter of how much time was spent in the past. Sometimes servers owners think that since reloading a plugin reduces the hook time to 0, that reloading a plugin improves performance, but that stems from their not understanding that the hook time is just a historical sum. A plugin with high hook time might have been causing lag a moment ago, or an ahour ago. It's very unlikely that reloading a plugin would actually improve the performance of that plugin. It would need to have uncommon design flaws for that to be the case. The results of your experiments may be due to coindicence or indirect causes.
The high hook time you are seeing for stacking related plugins (e.g., in Stack Size Controller) are caused by extremely high hook call frequency, not caused by the plugin. People have reported the same issues with Stack Modifier. If the issue went away when switching stack plugins, that is very likely a coincidence. Stack Modifier actually uses more hooks than Stack Size Controller, alhtough it doesn't use OnMaxStackable. Maybe whatever is triggering the hooks calls is calling OnMaxStackable more than other hooks, at least on your server.
I'm not trying to point fingers. I'm simply desperate and since using o.reload sends the server fps back up to normal, I assumed that to be an indication of causation. We've never had this bizarre problem, our server host can't pinpoint the cause either. On the other hand, very few of our plugins got updates with this facepunch update, and that is usually a sign for concern (or at least a place to start). It's a dedicated (actually dedicated, as in our own) server and last month everything was perfect. So something in the update on thursday broke... something. Pursuing that something is proving to be ... long.
Try this debug plugin to deal with the item related hook calls.
thank you, i will do so
restoreupondeath continues to be the standout for us. We fixed most of our earlier problems by removing some roaming bots per biome (botrespawn issue). But we keep gettign 1-second lag spikes, wher e the server fps drops suddenly to around 10-20 server fps before 1 second later going back to normal (around 55 server fps). In that second, the lag is horrible -- but for a very short spike. It's pissing people off if they happen to be doing something when that 1-second happens. But the only thing that's showing in the console is the server save point and restoreupondeath taking longer than anyhting else.
WhiteThunder
Try this debug plugin to deal with the item related hook calls.
Can he detect more hooks? I can pay