I just setup a new PBXact Saas server w/ the Solus Distro off the wiki.
All seems to be working well, still testing and playing with modules I didnt have in the past.
This is my first full adventure in hosted VoIP
My question is…
Would it be a functional and wise idea to setup software VPN (openswan) on each of the virtual servers, or would it be too intensive?
I am open to any input, suggestions.
Thanks in advance for the help.
The server specs are:
Dual Xeon L5520 (quad core 2.26ghz)
mirrored 512gb SSD
KVM Tempate pbxact cent 6.5 10.211.65_6
We do not recommend adding anything that is not already installed on your PBXact system. If issues arise on your system and custom items are installed support will inform you we can not provide assistance and it may effect the ability for us to upgrade if something you install has dependency on other versions of packages that we try to upgrade at a later date.
So you are running the Solus Controller as a VM on a node?
Also the SSD’s especially mirrored are a poor performance and longevity choice. I would have a 5 drive near line SAS or a 3 drive 15k SAS both with an additional hot spare.
We just brought up our 6th Solus node. 1 by 1 the blade server bays are being scrapped. We could not be happier.
Just a little plug for Schmooze, we are a PBXtended reseller. When we create a new PBXtended deployment the Schmooze portal creates in on the Solus node. This has worked since the platform was based on Proxmox. It is very slick.
That makes perfect sense. Hadn’t really thought about the ongoing updates - was just trying to think of the most efficient way of doing things.
We have a dedicated SolusVM Master. Also, thinking of using the master as a backup server, until we grow a few more nodes. Then we just brought up our first node, which is the server I listed the specs for.
I debated the SAS vs SSD - it seems its a never ending debate. We originally spec’ed the server out with SAS drives, but our sales rep said go SSD, that they rock for VM. I have never used SSD in a server, so I thought I would give them a try. Do they really hinder performance that bad? I know longevity is shortened, but was willing to take a shorter drive life, if they were better drives. Guess I lost out on both aspects.
We brought our first node and instance up on Tuesday, so far I love the setup and ease. Just trying to get everything in line before we go from test to production. I am really looking forward to working more with Schmooze, everyone I have ever talked to there is awesome, friendly and knowledgeable.
Any insights or tricks of the trade on running a successful PBXtended (been told its been rebranded to PBXact Saas?) setup on Solus?
I was unaware you could create the deployment in the portal and it would create it on the node. I created it then added it to the portal.
The only thing that I dont know if I did wrong or if it is how it is suppose to be is, my instance that I created shows the drive size as 3gb in the system status page and solus shows it as 25gb, is that normal?
THe debate over the SSD’s can go on all day.
Let’s table that if you are getting the performance you expect.
Yes, I still call it PBXtended old habits hard to break.
With regard to the volume error just exactly where are you talking? When you fired up the node did you provide the name of the LVM group of the volume you want to keep the KVM images on to the Solus node parameters?
What does the solus LVM display show for the node? What does lvdisplay show on the node?
SSD performance may deteriorate (as happens to rotational disks) but the matter, in this case, is that (SSD) wearing could not be controlled (from an high level point of view) by the Linux OS using usual TRIM/discard if the underlying physical devices (SSD disks) are “hidden behind” a software RAID made with md which, AFAIK, isn’t able to pass down to SSD disks in the array the TRIM command as instead happens when one or more SSD disks (with/whitout LVM on them, if I recall well) are “directly” managed by the OS (via Linux this happens setting discard at fstab level or scheduling a write cache flush via fstrim through a cron job or doing it manually from time to time). Probably the discussion about the correct usage/deployment of SSD Disks (single or in arrays) is more than that and depends greately on the implementation’s typology/scenario…and could last not only a day long…but forever.
Sorry for the delay getting back on here - the volume issue is fixed - I was having an issue that the VPS was going offline at random - I logged a ticket w/ Solus - mentioned the drive issue - they logged in and said all is well now, and everything works now. I must say Solus has prompt support also.
Solus is by far the best deal on the planet. I can’t say enough good about it.
If you give them good info, they are right on top of it.
We have been adding Solus Nodes at a dizzying rate. Turning up a 12 core 128G monster this month.
We only use KVM
We were using Proxmox prior and have not looked back.
Good to hear that its an ongoing great support and not just to get you hooked.
Those the normal nodes you use? How many instances you PBXact are you able to run on a box that size?
You run all your nodes in the same data center?
Thats all I have setup thus far is KVM - a couple cent boxes and couple pbxact - I pondered brinign up a windows server to see how it would run in KVM, any experience?