I think I fixed the console access from digitalocean. I don’t need it and kind of hate these cloud consoles, but for the next person, they’ll probably want it.
Memory appears to have stabilized at just under 60% usage per DO graph after maintenance this morning:
This is better, but still higher than I expect. I’d feel way better if we resized the droplet but this will double our monthly costs.
As I’ve said before, I don’t have much expertise in the infrastructure that we’re using to host the forum, but my take is to let this go unless/until we see some performance impacts on the site.
And to be a little flippant, what’s the worst that can happen if the system does choke? The forum is offline for a while–big deal. This forum isn’t essential and nobody’s getting fired over any technical glitches.
There’s also the financial aspect. If we are indeed seeing slowly shrinking member participation, I imagine that will (and maybe already has) impact our income level from one-time and ongoing donations. Seems like a bad time to sharply increase our costs.
All just imo, take it for what it’s worth. Thanks very much for taking on the admin role and being so conscientious about it!
2 Likes
yes I am probably the boy crying wolf here. DO is now showing 80% memory use again - but free -h seems to look fine.
$ free -h
total used free shared buff/cache available
Mem: 7.8Gi 2.6Gi 245Mi 3.1Gi 4.9Gi 1.8Gi
To answer your Q, in the case of a true OOM (out of memory) situation, I don’t want to find out. It may be likely I lose access to the VM. However, realistically - and I have no idea if this is true, just theorizing - the processes would crash long before that (the site would go down) and free up enough memory for me to take control of the VM and fix whatever is going on.
The thing is, I really don’t ever want to have to do this so it’s not going to happen.
The reality of this situation, honestly, is that we provision much larger servers to do much less at my current job. Eventually we will want to upgrade it, but there seems to be a fatalistic attitude about this site so idk.
I am wanting to do some stress testing of the site. I don’t know how to simulate that, but I could figure it out.
This sounds exactly like the discrepancies I used to see between top and glances, which makes me think the sudden spike might be the result of DO changing how they monitor rather than an actual increase.
1 Like
yea this is my conclusion as well. System memory use correctly shows 35% when I first ssh into it. Free looks fine too.
Deleted [redacted’s] account
Revoked Risky’s mod status per request
There appears to be a limit (30) to the number of people that can be added to a PM thread. Is this something that can be increased?
I’ve made it 500 for now until the first time we get a mass spammer.
4 Likes
thank you
I’m not going to be around for a few days and won’t be responding to any requests during that time.
That was very nice of you ggoreo
4 Likes
As part of our discussion about the transition to a new admin, jmakin and I have decided to rescind his admin powers over the software now, and reinstate them temporality if he needs admin access to do any of the server stuff in order to hand that off to the incoming admin.
5 Likes
Done
3 Likes
Thank you ggerorio
2 Likes
I made @spidercrab an admin. I do not think he plans on doing much if any adminning, but we need someone else with access to the admin features in case I die He is already the treasurer and is on the hosting account under his actual name and pays the bills, so he seemed like a good option.
4 Likes
you’re going to die?
Yes
1 Like