![](/static/66c60d9f/assets/icons/icon-96x96.png)
![](https://lemmy.ml/pictrs/image/ucPeLo62DS.png)
Are they roughly 55GB compressed?
Are they roughly 55GB compressed?
You suggested just adding the ISOs to local-lvm. Do you think it would be feasible to simply delete the local storage completely and then extend the local-lvm after, storing the ISOs there? I know extending volumes is much simpler than shrinking. And I imagine deleting completely is also easier than shrinking?
It’s really that much of a hassle to fiddle with the volume sizes?
Due to my understanding of it, I was hesitant to use AC recovery in the case that the power goes down more than once in a short period. It could drain the UPS to the point that it might not be able to sustain enough runtime for a proper shutdown. But I’m also a bit confused about the setup here. If the server is sent a signal to shutdown due to a grid outage, who is telling it the grid was restored? The server would always detect power because of the battery backup, so I don’t think AC Power Recovery would work in this case, no? I believe I have the UPS comm server (probably apcupsd) installed on the server itself, so there’s no way for it to know to wake up unless from an outside source.
Maybe you have some further incite into how to make that setup work properly.
I’m brainstorming here, but would it be possible/feasible to have the Unifi Dream Machine execute a script everytime it turns on telling the server idrac to power up. I’d have to see if the UDM has that ability as well. The UDM turning on would only really happen if power was restored from an outage. Otherwise I could send a command manually once I have access to the network.
I was under the impression you need a license to use remote access via idrac. Do you know if it also requires a second run of ethernet cable separate from the server NIC?
Very true. Traefik is pretty awesome. Thanks for the input!
I’ve had the IP for a couple of years so I can rule that out. The only thing different than usual is I recently added a backup server to the VPS network. I’ve been doing a remote backup from one server through the VPS to the backup server over the past week. It’s a 4TB backup averaging 4MBps.
My guess is that the VPS provider’s algorithm is bungling the bandwidth calculations, possibly refreshing the bandwidth amount incorrectly which to the black hole detector appears as a sudden spike in bandwidth rather than a steady flow.
I’m going to keep running the backup and compare how long after the backup starts that I get a black hole trigger. If it’s relatively consistent then that might be the problem.
I imagine that’s essentially what I’ve accomplished with Traefik already. The question I have is if Geoblocking does much to mitigate a DDoS. I know for sure it’s at least useful to block third world scammers and bots from running hacking scripts against my server.
I appreciate the tip but as a privacy minded self-hoster I try to avoid companies like cloudflare. Surely there has to be a way to diy DDoS protection?
You can connect your main server and backup server to a VPS with wireguard. The main server backs up proxmox vms and cts to a proxmox backup server on the backup server. Nextcloud data can be backed up with something like duplicati encrypted over sftp to the backup server. Only hiccup about backing up Nextcloud is you should put it into maintenance mode first. You can write a script pre duplicati backup and post backup to enable and disable maintenance mode.
Interesting. I hadn’t considered using a hardware key for SSH. I’m essentially using my desktop machine as a hardware key in a way, but obviously a dedicated hardware key would be best.
I’m curious if anybody has a more “self-hosted” solution, but I use burner numbers through the MySudo app and simply delete the number and buy a new one every few months.
If you look up Michael Bazell, he has a strategy for bulk buying voip numbers by tricking voip providers into believing you’re a large established business. You could buy dozens of numbers and just cycle through them. But that method requires a lot of work and social engineering and the providers are becoming privy to those tactics.
🕵️