Shulou(Shulou.com)06/03 Report--

The concept of "super convergence" can be simply understood as: merging virtual computing platform and storage together, forming a storage pool with the hard disk in each server, and providing IT infrastructure such as computing, network, security and storage needed by the data center in the form of virtualization. Here, we will also explain the advantages of hyperconvergence over traditional virtualization solutions.

Horizontal and vertical scalability

As the name implies, scale-out means adding servers as needed when storage and computing resources are found to be insufficient. For example, when the user's shared storage is full, the user has to spend a lot of money to buy a new storage cabinet, but the resource utilization of the storage cabinet is very low. On the other hand, users who use the super-convergence scheme only need to invest less to buy a new server to join the cluster and can expand the storage space.

Convenient provision of multiple copies to improve data security

The super-fusion scheme can easily support 2-3 copies. When some servers are damaged, if the super-fusion scheme is adopted, the required data will still be stored in the corresponding copy and the work can be carried out normally. Compared with the traditional shared storage, when users want to make two copies, they have to buy an identical storage device for backup, which increases a lot of IT investment.

Distributed storage, shortening the distance between computing and storage

The traditional shared storage needs to transmit data through network cable or optical fiber when reading and writing data. While the super-fused distributed storage reads the local copy data directly, reduces the time of data flow through the network line or optical fiber, and speeds up the data reading speed.

Integration of software and hardware to save money, effort and worry

The software and hardware integration supported by the super-convergence solution means that users can easily deploy the cloud data center at one time, including server, server virtualization, storage virtualization and other virtualization software. Compared with many users will buy hardware and software separately, the procurement cost is higher. At the same time, the software and hardware of the all-in-one machine has been implanted into the hardware when it leaves the factory, and it has passed the compatibility test, and the user can set it up directly to the computer room, turn on the power and simply configure it.

Summarize the characteristics and advantages of super integration:

Decentralization: there is no special control node, so there is no need to consider the single point problem of the control node.

Centralized storage: the traditional private cloud cluster architecture ensures that availability is achieved by shared storage. But shared storage itself is a single point, although it can improve availability in the way of multi-disk redundancy and dual controllers, there is still a dilemma that IO is centralized and performance can not be improved.

Lower construction cost: get rid of expensive centralized storage, disk distribution to local physical servers, the investment cost is greatly reduced.

Online capacity expansion: do not stop service, expand the capacity of physical machine accessories (memory, cpu, etc.), and even add physical nodes.

Brief introduction of PVE cluster building process: (3 nodes PVE1-PVE3:192.168.26.210-212)

1. Install and download the official ISO image of pve. It is easy to follow the prompts for the next step (if you have a special storage configuration, you can also configure it later, but it may be relatively cumbersome.)

After successful installation, use the browser WEB to log in:

2. Log in (through ssh) to update the source and system of three nodes.

Rm-f / etc/apt/sources.list.d/pve-enterprise.list

Echo "deb http://download.proxmox.com/debian/pve stretch pve-no-subscription" > / etc/apt/sources.list.d/pve-install-repo.list

Wget http://download.proxmox.com/debian/proxmox-ve-release-5.x.gpg-O / etc/apt/trusted.gpg.d/proxmox-ve-release-5.x.gpg

Apt update & & apt dist-upgrade

3. Start creating a cluster and execute root@pve1:~# pvecm create pve-cluster on pve1

4. Execute root@pve2:~# pvecm add 192.168.26.210 on pve2. Note that password authentication is required.

5. The command root@pve3:~# pvecm add 192.168.26.210 is also required for password authentication on the PVE3 node.

6. The cluster is successfully created to view the cluster status: any one of the three nodes is OK.

Root@pve1:~# pvecm status

Switch another node to try to check the status.

Root@pve2:~# pvecm status

7. Go back to the WEB console and find that all three nodes have appeared (any node in the centralization is fine)

8. Establish reliable storage for the cluster (Ceph storage, note that all three cluster nodes need to be executed)

Root@pve3:~# pveceph install-version luminous

9. Initialize the storage cluster network segment for communication. Note that you need to share the same network segment with the PVE cluster, otherwise an error will be reported (you only need to execute it on the PVE1 node once):

Root@pve1:~# pveceph init-- network 192.168.33.0

No active IP found for the requested ceph public network '192.168.33.0 pve1' 24' on node'

Replace it with the same network segment:

Root@pve1:~# pveceph init-- network 192.168.26.0

10. Create ceph cluster storage Mon monitoring (all three nodes are required, otherwise high availability will be invalid. If there is a node failure, it will stay stuck until the MON node is restored)

11. To create ceph cluster storage mgr (Creating Ceph Manager), you also need to install all three nodes or you will not be able to achieve high availability.

Pveceph createmgr # but the prompt already exists because the monitoring MON is installed automatically when it is installed.

12. To create Ceph OSDs (Creating Ceph OSDs), all three nodes need to perform the same operation.

13. Create a cluster storage resource pool (Creating Ceph Pools). This operation only needs to be performed on any node.

Root@pve1:~# ceph osd pool create pvepool 128 128

Check the storage cluster status ceph-s at this time.

Alarm: the solution of ceph:health_warn clock skew detected on mon

There are generally two reasons for the cluster status health_warn:clock skew detected on mon node, one is that the ntp server on the mon node is not started, and the other is that the time deviation threshold of mon set by ceph is relatively small. Investigation should also follow the first reason, and then the second reason.

Modify the time deviation threshold (remember to restart the service)

# vi ceph.conf

Add: under the global field:

Mon clock drift allowed = 2 mon clock drift warn backoff = 30

# ceph-s

Show health_ok for problem solving

14. Add RBD cluster storage (open a browser and log in to the WEB console configuration)

Note that you choose to use a super-converged CEPH pool.

View rbd cluster storage configuration information

15. Test creation on virtual server PVE1

Install CENTOS6.6

Pve3 install XP

Closing other nodes (Node 1) does not affect the installation operation

Test High availability: configure first

Test: PVE3 received an alarm after shutdown (email address was configured during installation)

Two minutes later, you will receive an email indicating that the failure has been successfully transferred.

The virtual server has been transferred

Take a look at the browsing status of the WEB platform:

We restore PVE3.

It is automatically migrated back after recovery because of its higher weight (direct and failback is not recommended in the production ring environment, which is configured as needed).

Clone virtual server in Proxmox

Migrate virtual servers (pve3 to pve1) in Proxmox

Configure spice analog terminal login (you can directly use the terminal to set the connection operation virtual machine to realize the cloud desktop):

Add something to the configuration file: args:-spice port=61001,disable-ticketing,seamless-migration=on

Starting the virtual server qm start 102 (which is started here with the command) can also be done through the WEB console.

After startup, it is found that one more port is open, and the operator server can be connected by configuring the terminal to connect to this port. Here, login is simulated by VirtViewer software.

Change the server and try again (linux):

You also need to modify the configuration file:

Root@pve3:~# cat / etc/pve/qemu-server/101.conf

Bootdisk: sata2

Cores: 1

Ide2: none,media=cdrom

Memory: 1024

Name: Centos6.6

Net0: e1000mm 26vmbr0: e1000mm 26vmbr0: 01v 01v 01v 01m m br0.

Numa: 0

Ostype: l26

Sata2: ceph:vm-101-disk-0,size=20G

Scsihw: virtio-scsi-pci

Smbios1: uuid=73100396-3804-4002-815b-4017d81199a9

Sockets: 1

Vmgenid: bb0d8fc0-a6c7-4283-81f3-6766c4f150a2

Args:-spice port=61002,disable-ticketing,seamless-migration=on

Save exit and start the virtual server

Connection effect:

It should be noted that the terminal is connected to the platform server IP plus port, which is a port-based application, and the virtual server configured after we connect to the server uses the IP address for the Internet, so don't mix it up. Even if we disable the network card, it can still be connected through the terminal, but the server cannot go to the external network (communicate with other servers) to see the effect in the picture, which is different from our usual remote access.

Cd

The functional effect is temporarily displayed here, and other functions and supplementary follow-up articles are further improved. I haven't written for a long time, and there are a lot of cursive articles.

Reference: https://pve.proxmox.com/pve-docs/chapter-pveceph.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.