„Telia Cloud Platform“ Vartotojo vadovas

„Telia Cloud Platform“ yra viešosios debesijos (angl. public cloud) paslauga ir šio vartotojo vadovo tikslas – padėti paslaugos klientams ja naudotis efektyviai ir suprantamai.
Šį vartotojo vadovą sudaro keli skyriai, kurie iliustratyviai paaiškina paslaugos sudedamąsias dalis ir pateikia pavyzdžius, kaip kiekvieną iš tų dalių pritaikyti savo poreikiams ir paruošti debesų kompiuterijos infrastuktūrą naudojimui.

Vartotojo vadovas


Linux VM Backup in OpenStack with Veeam Backup Agent

There’s a way to have user friendly backups in an OpenStack environment – interested? Then this post for you!

For quite some time now I’ve been thinking about backups in OpenStack – how  can I have a simple and clear solution to make a backup of Linux virtual machines. At the moment I can think of 1 or 2 backup solutions based on snapshots. These solutions are not exactly what I’d like or I could have confidence in the data safety. I don’t really like snapshot based backups so I’ve tried an alternative path of making backups in the old fashioned “Physical server” backup way.

I have selected a product from Veeam – Veeam Agent for Linux FREE version. It’s a free edition of their backup software – and I like free 🙂

Here’s the idea:

In this diagram you can see that one server in the infrastructure is going to serve as a backup storage to store Veeam created backup files. In this particular case the backup storage server will use Samba and CIFS. I think it will overall help me to improve the security side from Wannacry, Petya A or some other variants of crypto viruses. One could challenge a common CIFS shared folder? I say yes because it will be restricted to Veeam as a target repository to store backup archives. Another idea is to store backups in some external location if you’re really cautions 🙂

So, let’s start with the implementation.

Continue reading “Linux VM Backup in OpenStack with Veeam Backup Agent”

Intel Xeon 6154 (DL380) vs AMD EPYC 7551 (DL385) on HPE ProLiant Gen10

Our partners at HPE have generously provided us with a couple of ProLiant Gen10 servers, and I’m going to do some testing using Red Hat Enterprise Linux 7.4 with KVM. (Update: had to switch to CentOS 7.4 with RDO OpenStack later on).

I’m really excited to see HPE being a front runner with AMD and DL385s, as well as AMD finally releasing a CPU that’s finally (and hopefully) will provide a much needed competition for Intel in the datacentre space.

Here’s the gear I’ll be working with for the next couple of days:

1. HPE ProLiant DL380 Gen10

This one has two Intel Xeon Gold 6154 CPUs and 1.5 TB of RAM (768 GB per CPU socket) in 64GB LRDIMMs. BIOS updated to U30 v1.32 (02/01/2018).

Processor Name Intel(R) Xeon(R) Gold 6154 CPU @ 3.00GHz
Processor Status  OK
Processor Speed 3000 MHz
Execution Technology 18/18 cores; 36 threads
Memory Technology 64-bit Capable
Internal L1 cache 1152 KB
Internal L2 cache 18432 KB
Internal L3 cache 25344 KB

Networking is HPE Ethernet 10Gb 2-port 562FLR-SFP+ (Intel X710) and HPE Ethernet 10Gb 2-port 562SFP+ (Intel X710) for a total of 4×10 GbE ports, and HPE Ethernet 1Gb 4-port 331i (Broadcom BCM5719).

Storage is pretty basic as we have SAN infrastructure – HPE Smart Array E208i-a SR Gen10 with 2xVK000240GWJPD SSDs (240 GB) in a RAID1 configuration.

2. HPE ProLiant DL385 Gen10

This one has two AMD EPYC 7551 CPUs and 1.5 TB of RAM (768 GB per CPU socket) in 64GB LRDIMMs. BIOS updated to A40 v1.04 (12/12/2017).

Processor Name AMD EPYC 7551 32-Core Processor
Processor Status  OK
Processor Speed 2000 MHz
Execution Technology 32/32 cores; 64 threads
Memory Technology 64-bit Capable
Internal L1 cache 3072 KB
Internal L2 cache 16384 KB
Internal L3 cache 65536 KB

Networking is HPE Ethernet 10Gb 2-port 562FLR-SFP+ (Intel X710) and HPE Ethernet 10Gb 2-port 562SFP+ (Intel X710) for a total of 4×10 GbE ports, and HPE Ethernet 1Gb 4-port 331i (Broadcom BCM5719).

Storage is similar to the one above – HPE Smart Array E208i-a SR Gen10 with 2xVK000240GWEZB SSDs (240 GB) in a RAID1 configuration.

Continue reading “Intel Xeon 6154 (DL380) vs AMD EPYC 7551 (DL385) on HPE ProLiant Gen10”

Deploying OPNSense appliance on Telia Cloud Platform

It was announced earlier that there will be a post regarding the deployment of OPNSense VPN, LB and FW appliance on Telia Cloud Platform (TCP), so here it is 🙂

There are two ways to deploy it:

1) manually (create networks, subnets, spin up VM with proper networks);
2) via orchestration, using a Heat template.

In this post I’ll focus on the second option for the deployment process but the post-deployment configuration is the same for both options. So let’s start!

Continue reading “Deploying OPNSense appliance on Telia Cloud Platform”

Telia Cloud Platform story at Telia Technology Day 2018 [LT]

In case you’ve missed it or just want a recap, here’s the story about the Telia Cloud Platform’s journey to date, lessons learned and plans for the future, presented at Telia Technology Day 2018 (in Lithuanian).

Storage performance in Telia Cloud Platform

This is going to be quick one because a picture is better than a thousand words.

We have three storage tiers in Telia Cloud Platform – Tier-1 on SSD, Tier-2 on SAS and Tier-3 on NL-SAS, and below are the results from CrystalDiskMark 6.0.0 in a Windows Server 2016 VM on a hpc1.a5 flavor.

As for the IOPS@4KB Q32T1 (click on the links below for a full CrystalDiskMark output):

Telia Cloud Platform test results!

We’ve run a series of tests to provide our customers with a reference as to what can be expected from the platform flavors.

These tests span across a number of GC1 (B3, B7, B11, B17) and HPC1 (A3, A7, A11, A17) flavors. In a nutshell, we demonstrate the vCPU, RAM and combined system performance for 2, 4, 8 and 16 vCPUs with 4 GB, 16 GB, 32 GB and 128 GB of RAM respectively.

We’ve used Phoronix testing suite to run a series of tests on CentOS Linux 7, Kernel: 3.10.0-693.17.1.el7.x86_64 (x86_64), Compiler: GCC 4.8.5 20150623,
File-System: xfs, Screen Resolution: 1024×768, System Layer: KVM.

The tests were run on a live system in a production environment, i.e. no fake or demo lab setups, and artificially boosted numbers.

The tests are completed, a combined output is available here: Telia Cloud Testing Results.

So you’ve got your OpenStack project in Telia Cloud – now what?

Once you get on board the Telia Cloud Platform you essentially get a virtual data centre, and at the starting point it is empty with nothing except a pre-provisioned network for the Internet. Of course, there are certain resource limits imposed but these are quite liberal and should fit most needs, but you can request an uplift if needed by raising a support ticket with service desk or billing self-service.

So what can you do now? Naturally, there might be an urge to jump in and spin up some VMs, create volumes of different storage tiers and do some testing, and that’s absolutely fine. In a cloud environment you’ve got the power to have your own private networks, virtual routers and virtual network appliances that have a very significant impact on how you will consume all that virtual infrastructure available to you. The infrastructure design considerations and decisions are no different to those you’d make with physical infrastructure or plain virtualization. Basically, on the infrastructure level you’re dealing with three main areas: compute, storage and network.

Continue reading...


The compute area deals with the resources for your basic VM needs: vCPUs, RAM and root disk size – all three are defined by a concept called “Flavor”. This flavor defines the “size” of your VM, and in our platform there’s a great number of them, plenty to choose from. In a cloud environment you’re no longer able to create VMs with very specific sizing parameters (for whatever reason, like 3 vCPUs and 5.7 GB of RAM) – you have to choose from a list that the service provider gives you. We tried to be creative and came up with a pretty long list that cover most use cases.

Then you need to choose the source for the operating system. We are providing a number of OSes to choose from. If you’re not happy with the selection – feel free to ask for additions or upload your own image. Bear in mind, that you’re responsible for the licensing and intellectual property violations if you’re using your own images. One very important aspect about the images – we neither know what the root or Administrator password is, nor we can somehow hack it for you, sorry. The Linux-based images can be initially accessed using a key-based SSH authentication only, Windows-based instances ask the user to set the Administrator password upon the first boot – make sure you check the console tab of your VM in the self-service portal. Once you have your Linux VM running – you can then set the root password you want, snapshot the VM and then use that snapshot to create additional VMs with a pre-defined root password that you know, but we endorse and promote a more secure SSH key-based authentication as a default and initial option.

The root disk is an important aspect – this is the location for the operating system and the applications that you might want to install. In a Windows world it’s disk C: 🙂 This kind of setup might seem to be restrictive but only at the first glance. The idea behind this root disk concept is that the application or whatever other data should resize in it’s own dedicated space and that space is provided by storage volumes (or LUNs, in the old school speech).


In our platform we have three tiers of storage to choose from: Tier 1 – SSD storage, Tier 2 – SAS storage, Tier 3 – NL-SAS storage. It’s entirely up to you to decide what kind of storage you need for your business. Can you also add volumes from different storage tiers to same VM and aggregate that storage using LVM in Linux or Storage spaces in Windows. Windows can also leverage Dynamic disks but I’d strongly recommend Storage spaces as it’s a native Windows Server component since 2012 and much more flexible.

Once you create a storage volume you can then attach it to a VM. At the moment you can attach it to a single VM only – once we launch a new Telia Cloud Platform version (v2) you’ll be able to attach it to multiple VMs but not at the moment. The nice thing with volumes is that you can create snapshots of them (think – instant backup) and even create new volumes from these snapshots if needed.

When creating a new volume from snapshot you have to use the same tier of storage (which may seem restrictive but that’s the nature of snapshots). However, you can then safely change the tier of storage by retyping the newly created volume from snapshot to whatever storage tier you want – select the action “Change volume type”, then set the name, target storage tier and make sure you select migration policy “On demand” because the policy “Never” is, well, never…

You can also extend volumes to make them bigger. Naturally, you can’t do this though with a volume that’s in use and attached to a VM because it’s risky to do so. So if you need to extend the volume to make it bigger – make sure you safely unmount it in the OS and detach it from the VM before attempting to extend it. Or, alternatively, you can use snapshots and creating new volume from snapshots to make this happen and then sync the new volume with the original volume in the OS whilst both volumes are attached, and umount and detach the original volume when no longer needed. This is just one example, you can be much more creative than I am 🙂


Network is absolutely fundamental and planning your network properly is a must. We don’t know your business that well as you do, hence we don’t pre-provision any private networks for you. The only pre-provisioned network that you’ll find is “Public-Internet” which is, well, The Internet. You get your Internet accessible IP addresses from that network.

While you can, of source, plug your VM directly to the Public-Internet network (think – WAN), it may not be the wisest thing to do, unless you really know what you’re doing. Some old school people get confused about virtual or floating IP addresses because they like to see the IP address in the NIC configuration. With virtual/floating IPs it’s different – you see your private network IP address on the NIC and assign a virtual/floating IP address on the cloud platform level. Simply put, the floating IP is a 1:1 BI-NAT operating on the virtual router level and that’s the default operating model in any cloud environment. To use this kind of model you need to do two things – 1) create a new project internal network (think – LAN) with a subnet in it; 2) create a virtual router with a gateway set to Public-Internet and add additional interface to your project internal network. By doing this you’ll be able to launch your VMs in your private project network and provide them with the ability to SNAT to the Internet (get updates, download scripts and what not) and assign floating IP addresses only to those VMs that really need to be reached directly from the Internet by DNAT. The nice thing about floating IPs is that they can be viewed as IP address stickers – such “sticker” can move around servers very easily without any need to reconfigure your NICs on the VMs.

All in all, I recommend setting up at least one project internal network with a virtual router connecting it to the Internet and use floating IP addresses for the use cases requiring access from the Internet to the VM provided resources. Oh, and don’t forget about the Security groups – they are your endpoint firewalls, and you can have multiple security groups, each tailored for every specific VM role, like web servers, DB servers, middleware servers, etc. The rule set in the default security group permits only outgoing requests (egress rules). If you need remote access, like SSH, make sure you add rules permitting incoming requests (ingress rules) and don’t leave that default CIDR – every address on the Internet matches this rule, try to be more specific where possible and reasonable.

There’s also another option for your networking needs – virtual router, firewall, load balancer and VPN appliance. We’re happy to provide OPNsense virtual appliances, more on this subject in the future posts.

Close post...