-
Notifications
You must be signed in to change notification settings - Fork 109
Description
What would you like to share?
There is a very frustrating delay on my self-hosted instance (single user) wherein I have to wait at LEAST 30 seconds to get a results page. It can go as high as 45 seconds at times. It just makes the whole experience just really poor and is a deal breaker for me.
I see #615 and I checked the line 141 in file src/server/routes/search.rs
and since I am on the rolling
branch, it has the recommended change.
However, I see NO improvement whatsoever. And what is described in #615 is not resolved, at least on my box.
Do you want to work on this issue?
None
Additional information
How I build:
Clone the "default" branch which is rolling.
I sometimes do cargo clean
(which is probably not needed).
I have read the instructions here and for my latest build, have defined the variable as such:
export PKG_ENV="dev"
However, this does not seem to have an impact as defining like this: export PKG_ENV="prod"
before rebuilding does not remove the delay.
I have tried to build with cargo build -r
or cargo build -r --no-default-features --features redis-cache
and this changes nothing re. the delay: it does change the speed for cached pages but a single user does not actually benefit that much from cached pages as one does not often repeat the same searches (however if one does, it instantly displays the cached page which is nice.)
I have tried to set debug = false
or debug = true
in config.lua
and no changes.
Same for production_use = false
or production_use = true
, still no changes.
What am I doing wrong here? Most search engines, including searxng
are MUCH faster than the websurfx
instance I run and build as described above. I suspect there is a problem with the intentional delay as referred to in #615 and in the documentation but cannot put my finger on it.