I solved this - problem was, after I changed port on MC2, I needed to download new MeshAgents that are setup for that port and install them to monitored PCs

@JaredBusch Thanks for the link. Good reading!

@JaredBusch I have a couple clients using Synology for their auth needs and it's been working extremely well.

Exported, cleaned up references to mongo and all better. No Mongo.

Export the database (this uses the running database, so I had to have Mongo fixed)

cd /opt/meshcentral node node_modules/meshcentral --dbexport"

Go in the config file and remove the "mongodb" line(s) in settings

nano /opt/meshcentral/meshcentral-data/config.json

Restart meshcentral

systemctl restart meshcentral # or /opt/meshcentral/mcstop /opt/meshcentral/mcstart

Import the database

cd /opt/meshcentral node node_modules/meshcentral --dbimport"

Restart meshcentral

systemctl restart meshcentral # or /opt/meshcentral/mcstop /opt/meshcentral/mcstart

If this is a systemd distro, you can just create a timer and service. That way instead of sending a mail message, it's automatically part of the logging for that unit.

If it's a recent version of systemd, you can have a user based systemd that doesn't need escalated privileges. I just finished setting up borg for backup since I'm on Fedora Silverblue now (couldn't use the old way), and I just set up user based systemd units to do the backup.

@Pete-S said in Checksum verification:

Did the piping example I posted above satisfy your switch requirement even though it isn't a switch?

Or maybe you're looking for something else?

Was busy dealing with the firmware update itself.
Yes. That works simply enough. Still annoyed that they don't have a switch..

@Pete-S said in Service provider - monitoring questions:

Thanks guys for your replies!

I had a feeling that monitoring was "underutilized" compared to what is technically possible. But as always, it's the business needs and the effort (cost) that determines the service level.

totally underutilized compared to what is possible. you are 100% spot on with that.

@scottalanmiller said in FortiGate - Speedtest:

@notverypunny said in FortiGate - Speedtest:

You should be OK with a laptop for these speeds.

A laptop, yes. That's not the point. The issue is something with this particular setup, laptop or whatever, can't get the speeds so testing with it doesn't tell him anything useful.

I've run into issues specifically for laptops where I couldn't get decent speeds, something either with the HW specifically or power-management / network stack wouldn't allow me to get the same speeds as a server / workstation connected to the exact same port / equipment. Regardless, you're correct that this laptop seems to have issues.

Is the package 100MB up and down? Is it a full 100 or a "burstable" setup (not sure how common this one is, but at a former gig we had a line that was provisioned at one speed but would allow temporary surges to another tier before automatically throttling back to the base rate.)

using iperf3
DOWNLOAD SPEEDS

iperf3 -c iperf.he.net -R -P 10

UPLOAD SPEEDS

iperf3 -c iperf.he.net -P 10

You can check a list of public iperf servers :
https://iperf.fr/iperf-servers.php
if you're in another region or if the HE ones are too busy

It really depends. If you don't do software transcoding even hundreds of call on quite low power hardware would be fine (say 2GB ram and 4 CPU cores).

We've found that using a Sangoma E1 card the same hardware would handle the double of the calls, only because Asterisk bases its timing source on the Sangoma hardware instead of using the dummy one.

One test that you can do is to run the command dahdi_test, the values should be not less than 99.996% for a good audio quality, specially for music on hold and conference rooms.

There are some hardware timers that do the same job for a very low price, ie

https://www.thedebugstore.com/asterisk-pbx-system-timer-cards/

@scottalanmiller said in Your Top 5 Azure or O365 tasks:

Number One Task for Us: Creating a New User, Assigning a License, Adding to Groups

Ditto.

You need to setup the EC2 instance with a WAN Dynamic IP (Elastic IP address). Then just go through the Internet gateway and have the security group blocking incoming traffic over the Dynamic IP provided to the EC2 instance.

@JaredBusch said in Install MeshCentral2 on Fedora 29 with MongoDB:

@scottalanmiller said in Install MeshCentral2 on Fedora 29 with MongoDB:

@bnrstnr said in Install MeshCentral2 on Fedora 29 with MongoDB:

Is it possible to migrate from MongoDB to NeDB after we've been running for a while? Easier to just start over?

Probably, but never considered doing that!

I think @Ylian mentioned that is was possible or was going to be possible. in the main thread.. not 100% sure.

Should be, they share a format. I've just never ingested a backup into NeDB before.

@Obsolesce said in Revisiting ZFS and FreeNAS in 2019:

@scottalanmiller said in Revisiting ZFS and FreeNAS in 2019:

Stuck supporting a crappy FreeNAS deployment today. What idiot installed this.

Was it still running or was that the reason you were called?

It was the reason that I was called 😉

@scottalanmiller said in I left Samsung for Huawei:

@Doyler3000 said in I left Samsung for Huawei:

When my son finds a way to destroy this one I'll definitely be looking at Huawei again, if they're not completely crippled by the Google stuff.

Great feedback. I've been doing some SAMIT videos on YouTube doing analysis of the Google situation. Personally, I think Huawei is going to weather the storm. But it'll be tough.

If anyone can it's them. Interesting piece of human history going on right now. I'm not sure how it will end. Maybe just sanctions being lifted

Also we could achieve hyperconvergence "the other way" (unlike having a global shared filesystem like Gluster or Ceph) but use fine grained replication (per VM/VM disk). That's really interesting (data locality, tiering, thin pro etc.). Obviously, we'll collaborate to see how to integrate this in our stack 🙂

63410266-b777-4ebd-83be-e8b7bdaa4816-image.png