@stacksofplates Thanks, good info. Im learning ansible now.

Here's what the alerts look like in both alertmanager and prometheus:

0_1499526971615_alertmanager.png

0_1499526982585_alerts.png

@obsolesce said in Systemd timers instead of cron:

@stacksofplates said in Systemd timers instead of cron:

Now just enable each one
systemctl enable backup-tower backup-pbx backup.target backup.timer

I'm not sure how this works.
Why wouldn't you just enable the backup.timer? Why add everything else in there?
I'm not clear on how it's all linked.

Yeah you don't. I honestly don't remember why I had it this way. Could have just been an accident. Was quite a while ago.

@scottalanmiller said in DigitalOcean Offers Monitoring:

@stacksofplates said in DigitalOcean Offers Monitoring:

@scottalanmiller said in DigitalOcean Offers Monitoring:

Day late on that one 🙂

I just saw the email this morning.

@stacksofplates said in DigitalOcean Offers Monitoring:

@scottalanmiller said in DigitalOcean Offers Monitoring:

Day late on that one 🙂

I just saw the email this morning.

We had a thread yesterday.

Missed it. Been busy at work.

Edgerouters can now handle custom dynamic dns providers without the need of the script - here is the guide with Cloudflare as the example: https://help.ubnt.com/hc/en-us/articles/204976324

But you learned something along the way. 😄

systemd has got this since early development. this, among other things (PID 1 master process...etc), was one of the reasons Poettering designed systemd. It borrowed this idea from other systems like this

So I also realized that if you have even a small number of records, a dictionary will become super long. So you can compact the dictionary like this:

records: ns1: {forward: 10.1.30.11, type: A, last: 11, rev: 30.1.10.in-addr.arpa.} ansible: {forward: 10.1.30.5, type: A, last: 5, rev: 30.1.10.in-addr.arpa.}

So while the other way may be easier to read, this saves a TON of space.

We are using this, by the way. Just deployed one this past week or so.

Trying to figure out if I should learn Ansible or Salt Stack.... hm..... Thoughts?

I am studying for my RHCSA... so maybe Ansible makes since?

I guess I could do a wildcard at the beginning and end of whatever I'm searching for with killall but that freaks me out a little. I could accidentally do a lot of damage with that.

So if it doesn't provide privilege escalation and only allows you to access processes which you are running, then how is it really any different from any normal hooking? So yes, it's unpatchable in the same way deleting all the files in your own home directory in Unix is unpatchable.

Looking at the code as well... we were doing this stuff years ago to see what was being sent wrapped in SSL for reverse engineering purposes, because tcpdump or whatever couldn't read it.

@scottalanmiller said in Red Hat Open Sources Ansible Galaxy:

That's awesome. Now just waiting on Tower, too.

Thought exactly the same: Free/open Tower would be awesome.

So I found a quirk, not sure if it's an actual bug. For incremental backups to work you have to use the NFS syntax (BACKUP_URL="nfs://<server>"). If you mount the share, and then write to the local mount it won't do incremental.