KernelCare and LibCare in a VPS

nfn

Hi

One of my VPS is relatively busy and I was thinking about using KernelCare to avoid reboots an receive patches.

From my understanding, LibCare should be added to the basket too because KernelCare only updates kernel and libraries updates need reboots too.

Are you using KC and LC in a VPS? Is LC needed?
What are your opinion about using KC and LC in s VPS?

Thanks

Boogeyman

https://github.com/cloudlinux/libcare

RTFM

yoursunny

You can buy KernelCare and LibCare, but the host node would reboot and ruin your uptime.
Hence, design a distributed system that spans two servers, so that either server can go down without affecting application.

nfn

@yoursunny said:
design a distributed system that spans two servers, so that either server can go down without affecting application.

This is something I've had in mind for a few weeks now, and since the most important and busy applications are developed by me, the effort to make them multi-server is low.

Furthermore, two of them are already distributed worldwide using 5 VPS.

I'm using CF DNS and HetrixTools for monitoring. Any suggestions for automated DNS change when the primary server goes down?

Kassem

@nfn said: I'm using CF DNS and HetrixTools for monitoring. Any suggestions for automated DNS change when the primary server goes down?

CF has load balancing solution from $5/mo. If you don't want that, you can simply setup a webhook notification in HetrixTools and call CF API to update the record when primary is down. Obviously your webhook link configured in HT shouldn't be on the primary server.

https://www.cloudflare.com/load-balancing/
https://docs.hetrixtools.com/uptime-monitoring-webhook-notifications/
https://api.cloudflare.com/#dns-records-for-a-zone-update-dns-record

AXYZE

@nfn said:

@yoursunny said:
design a distributed system that spans two servers, so that either server can go down without affecting application.

This is something I've had in mind for a few weeks now, and since the most important and busy applications are developed by me, the effort to make them multi-server is low.

Furthermore, two of them are already distributed worldwide using 5 VPS.

I'm using CF DNS and HetrixTools for monitoring. Any suggestions for automated DNS change when the primary server goes down?

Cloudflare Workers (very cheap or free, depending on how many requests you get)

Example
https://medium.com/@theotow/how-to-build-an-application-loadbalancer-for-1-1mio-requests-5c268e1d9d02
you need add function that will health check the servers and replace them, @Kassem idea is good starting point. Workers will be way cheaper than normal load balancing.

This has everything you need in one, but its on outdated API so you need to update it
https://github.com/wojons/cloudflare-lb