I just realized something Jeremy said. This is happening because it's acting as if the tiddlers have been lazy-loaded when they actually haven't been. So ... If you use the lazy-load option:
root-tiddler=$:/core/save/lazy-all then the Wiki will load fine without crashing. So if you really want to load this set, you can do so. -- Mark On Tuesday, June 4, 2019 at 9:58:20 AM UTC-7, solex wrote: > > Mark, > > I am not sure the exact number matters much. > So far anyone who tried to reproduce it with 2000 tiddlers got the server > to crash, so the issue is reproducible on most systems. > I would expect the exact number to differ depending on the available > memory and processing power, but the gist of the problem is that the system > cannot handle a few thousands of skinny tiddlers while being quite capable > of handling tens of thousands of "normal" tiddlers. > > On Tuesday, June 4, 2019 at 7:05:51 PM UTC+3, Mark S. wrote: >> >> I'm not sure what the "1800" tiddlers is based on. In my tests, all the >> tiddler titles loaded. Behind the scenes, apparently the server was stilll >> fishing for tiddler text. It must have been using an internal array to save >> results while waiting for (non-existent) data to load. The server crashed >> after the server had reported 1862 load attempts. But that's just as far as >> the reporting got. Internally, it may have had all the load events queued. >> The output reporting would have just lagged. So, I think,it would be more >> about machine processing maching. I'm running at 3.2 Ghz. I suspect solex's >> processing speed is similar. >> >> The way I captured the report is to use the > operator to funnel the >> output of the node launch into a text file. Then I could look at the file >> and count how many load entries there were. solex could repeat the process >> with his own device. It would be a doubtful coincidence if our count was >> exactly the same. In that case, maybe there's a pre-set number of threads? >> >> >> -- Mark >> >> On Tuesday, June 4, 2019 at 7:51:40 AM UTC-7, @TiddlyTweeter wrote: >>> >>> Jeremy Ruston wrote: >>>> >>>> This isn’t actually a bug, it’s more of an unexpected consequence of >>>> the way that lazy loading is implemented: in client-server mode, if the >>>> browser references a skinny tiddler (one that lacks a text field) then it >>>> requests the body from the server. >>>> >>> >>> My impression reading the thread is that the "break-down" occurs at a >>> specific point? >>> >>> I mean, a server overload I can't see happening at a specific number on >>> different machines. That has to be code? >>> >>> I remain hazy whether this is able to be replicated at a set number? But >>> that was my naive first impression--that it does? >>> >>> Side thoughts >>> Josiah >>> >>> >>> >> -- You received this message because you are subscribed to the Google Groups "TiddlyWiki" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/tiddlywiki. To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/e5f8b237-9368-47c9-bf54-d95bc7395ea2%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
