We been using FreePBX for maybe 7 years now. We have installed it in multiple client sites. Most of those seem to work fine. But our OWN system here at out office… is one constant Fn Babysitting job day after day. EVERY day it seems like there is some issue that disrupts our communications and I need to spend time tinkering with it to fix it and often no changes were made and all of a sudden it start or stops working.
FreePBX is running as a VM on ProxMox. Before this it was a VM onTrueNAS
On Ver 16 we were having this same problem. SO I dumped it completely and did a fresh install with 17. Same type of problems. But everything else on our network works fine. Web servers, mail servers, database servers
Today, for NO apparent reason, I can not register my soft phone on my cell phone. yesterday I was making calls with it, today 20 minutes Fn with it so far and cant get it to work. restart the phone, restart FreePBX, try to do update with Freepbx and it throws an error “Can not connect to http://mirror.freepbx.org/” Nothing changed on the firewall. nothing changed on our network.
Some days inbound calls work fine, I can receive calls on the desk or cell. Monday, when calling out the cell phone worked fine, but if i called out from the desk phone the person I called could not hear me. Fd around another 20 minutes trying to figure out what the problem is, never did… restarted a few times and it seems to fix itself.
This is so damn irritating, every time i need to accomplish something the first thing i need to do is F around with FreePBX to make sure it is working before I can accomplish anything any more. I am really tempted to call the phone company up and jsut go back to POTS lines.
Oh.. but our 7 client sites we installed at work fine.. one is processing 3,000 calls a month - but cant fix my own shit
I just restarted again. Could not get it to work. 5 minutes later it just started working. I was able to use call from my desk to the soft phone extension.
Great, I go to make a call to a client, and now it will not register again and I cant make the calls.
I just wanna bang my head against the wall. Actually, I might just do that. I JUST discovered that my cell phone has been attempting to use the WIFI connection instead of trying to use the cell data to connect to our server. So yea the registration is failing because the IP address the softphone is using is our external IP, not the internal LAN. AARrgggg… Ok, one problem Identified.
the problem here, despite what the fanbois will deny, is the VM. It has never been recommended to run a PBX in these things by those who know.
Get yourself a small cheap micro PC, lenovo think centre’s M73’s or later variant of the miniature size, ours uses avg 11-13 watts with odd peak on 23w if thrashing an SSD for say nightly backups, but only briefly, they purr along like a new born kitten never missing a beat.
You’re conflating KVM with containers. You can use KVM with no issue, issues happen when using containers.
But since threads like this generally give a summary complaint but never any real details such as configs, versions of things, what the actual troubleshooting was or any sort of logs or output for anyone to review and look at it’s hard to know what was the actual issue. Specially when in the ranting or subsequent posts it starts to come out part of the issues are PEBCAK it makes one wonder how many of the other issues are actually PEBCAK.
KISS - use a dedicated 64-bit pc with 4GB Ram and 250GB HDD or SSD
Take one of those from inventory - you have many to pick from in house, and put it to good use.
Grandpa says: Just because you can, doesn’t mean you should.
We have 75 hosted PBXs running on our vmware cluster. Have been doing so for over 15 years. As one of “those who know” I can assure you running Asterisk on a VM is not a problem if it’s configured correctly.
Same. Started with IncrediblePBX deployed on on-prem VMware in 2018. Then in 2020 moved to AWS FreePBX provided by The Web Machine. It’s been smooth and allowed peace of mind the past 7 years. Roughly 200 extensions and 5 physical sites.
@miken32 Can I pick your brain ? I need to talk with someone that has that type of use experience. I am looking to do the same here with our site instead of setting up more hardware at clients sites. Same, We have had FreePBX running for years as VMs, but something seems to have changed recently. We are in Hawaii and seems like a lot of issues we run into for all sorts of various things (even outside of FreePBX), the answer often turns into “its Hawaii.” A ship passing between islands a few years ago snagged and cut the undersea fiber cables. Killed most of the states internet for almost a week. An earthquake on another island knocked out power for our entire island here for 3 days. Almost a million people with no power. Cascade utility power failure. Took them 3 days to get all of the plants back online.
We have one client that noticed any calls going out to T Mobile numbers may not work for the person they are calling. Inbound calls seem to work fine.
@miken32 Can I chat with you directly ? For your PBX’s, Assuming you have multiple external IP addresses, one for each of the VMs ? Do you have any other infrastructure running on the same network ? web servers, email servers, etc ? Or is that network dedicated to phone PBXs ?
Speaking out of turn here. But without tangible specifics (e.g. - screenshots, log snippets, etc.) your thread hasn’t elicited too much of a response.
When it comes to outbound calls to T-Mobile customers, I assume T-Mobile employs STIR/SHAKEN technology, but from their standpoint I can’t see them blocking a DID from their network unless it was previously blacklisted somewhere. Depending on how long your organization has had the specific DID, that might be a thing to check. Otherwise, it is likely a networking or configuration issue on the FreePBX end.
When you say “it’s Hawaii,” if there have been power issues, Internet connectivity issues, PSTN issues, etc. that can impact operations, then this can be a factor. Aside from FreePBX troubleshooting, has anyone looked into your problem site’s overall network health? ISP latency, packet loss, routing hops, etc. If the ISP is providing sub-par service, than any services that require it can suffer.
If it were me, I’d purchase a basic Starlink package for the problem site. Connect a vanilla FreePBX instance to that, and test out with a very basic config. Maybe just one or two desk phones. Might sound like a reach, but for sites with a history of Internet connectivity issues it’s good to have a backup carrier for limited operations if something goes down. Our hardware firewalls have fiber ADI primary, CMTS secondary, and even Starlink as tertiary. Going down the ladder perhaps means less users can reliably hit the system, but it is for contigency purposes.
Or something else. Not having a debug of this issue makes this a guessing game. The fact that inbound works but outbound doesn’t isn’t an indicator that this is a network problem on the FreePBX end. Could it be a configuration issue, it can be but it could also be an issue with what is being sent in the INVITE to the upstream or something on the terminating side.
We need to see some real data/debugs on this T-Mobile issue.
Each of our PBXs has three interfaces: public, management, voice (internal b/w PBX and gateway). We only peak at 60-80 concurrent calls; on the 10 GbE links within our LAN we’ve found no need for QoS or anything. Voice isn’t our primary business so the VM cluster (couple of Dell R750s with 512GB RAM and a couple of 24-core Xeon CPUs each) has a lot of other things going on. But the PBXs are very light, we vary resources depending on number of endpoints, but typically 2-4 GB RAM, 1 CPU core, 16-32 GB disk each. CPU load is barely noticeable for PBXs these days. It’s a much different story than 15 years ago.
I think at this point, the internal infrastructure would only be one consideration. If @HawaiianHopeOrg thinking about mimicking what you are doing, i.e. hosting all the PBX systems themselves in a VM cluster they need to understand all things that need to be involved. The biggest being IP space because getting a /24 or more of IPv4 space is going to be a real PITA these days and could take over a year to get an assignment. IPv6 space would be gotten almost immediately. There’s probably a data center involved in your setup which means there is redundant power and connectivity. Most like a generator backup to said power.
So there’s way more to this than just building a VM cluster and hosting it in the back room of your office. I mean, if you’re actually looking to do it properly that is. Otherwise the back room is just fine.
Correct on all points (though we moved to a new NOC hosted by a mid-sized local ISP last year and got a /25 no problem within a week or 2.)
It’s definitely an expense that needs a lot of planning and consideration, but if you’re hosting anything client-facing you have to have that redundancy in place, otherwise one outage can cost you the business. Our hardware is overkill, but if one of our hosts (or switches or SAN cards or firewalls) fails, it’s all doubled up so the working device can handle the load solo until a replacement is installed.
Our only vulnerable point is we bought the cheap version of VMWare so we have to move our PBXs manually in the event of a host failure. It wakes me up at night sometimes lol
been there, done that!