It says Redis is in the stack. I wonder if something is wrong at that layer.

· · Web · 1 · 0 · 2

@dave I just ran `redis-cli flushall` and it didn't make a difference.

@theDanielJLewis Sounds like it's time for a chat with SpinupWP support. 🙂

@dave Yeah, I reached out to them. I might have to pay for priority support.

@theDanielJLewis That sucks man. Sorry about that. Cloud stuff can make you feel helpless at times. Just a black box.

@dave Get this. Vultr copy from LA to LA and retesting shows lower memory usage! But same loading time.

NJ copy is still loading.

@theDanielJLewis @dave

I bet it's DNS lookups. They always bite me in the butt.

Otherwise: Is the database on the same server as webserver?

@martin @dave It is _exactly_ the same. I cloned the fast low-RAM site exactly to the other server.

I can see DNS or server location affecting load time, but not RAM usage.

@theDanielJLewis @martin Are you able to shell into it, or are you just going by what's in the panel?

@theDanielJLewis @martin What's using the RAM? 100MB doesn't sound unreasonable for a WP site with the DB running locally.

@dave @martin Maybe, but keep in mind that the SAME site on a different server is using 1/4 the RAM.

Besides, _this_ particular site isn't plugin-heavy.

I have Query Monitor, but I can't find any way in it or any other tool that will show me _what_ is using the RAM, only how much is being used in total.

@dave @martin I do. But again, it only shows me processes, not plugins or what's causing those processes to use so much.

@theDanielJLewis @dave

The same site on different servers (if the config is different) can easily take different amount of RAM or CPU. Maybe the live server has PHP compiling enabled etc. - But I still think the load time in this case is probably DNS lookups. (Unless it's something like a really slow database connection).

@martin @dave SAME server config between staging and live.

And exact copy of live server to new instances is resulting in less RAM usage on the copies.

This isn't a popular site, either. Other sites on the server _are_ popular, but that shouldn't affect _this_ site's RAM usage, right?

@martin @theDanielJLewis gree

Same config, completely different everything else (location, hardware, routing, etc.)

West coast U.S. can really suffer from DNS latency depending on how the provider is bottlenecked in their routing.

Post a screenshot of your htop sorted by memory usage (SHIFT+M)

@dave @martin Same _kind_ or package of hardware.

After fresh reboots of both, live server on left, same-location clone on right.

@dave @martin MySQL is obviously using more resources on the live server. Under all those MySQL instances is this:

@dave @martin Bleh. I'm out of time to work on this right now.

I'll torture myself with it more later tonight.

@martin @dave My NJ instance just finished cloning. It loads faster (I'm near Cincinnati) and has same low RAM as my LA clone. Both 1/4 the RAM usage of the live server in LA.

Sign in to participate in the conversation
PodcastIndex Social

Intended for all stake holders of podcasting who are interested in improving the eco system