VMware vSphere5.5 Connection md3200i details (ii)

Adding iSCSI devices is simple, the properties of Vmbaxx under Host-"configuration-" Storage adapter-"iSCSI Softwareadapter

In the open window click on dynamic discovery, iSCSI server filled in 192.168.100.10, to ensure that the IP and VMware host can communicate, md3200i has 8 IP addresses, when you add, will be in the static found inside the md3200i all 8 IP. If not all of the IP can communicate with the VMware host, do not worry, md3200i himself will make a judgment, the formation of the path will only show the path can be successfully connected.

After adding the completion will prompt whether to scan the disk, if in the previous md3200i by copying the value of VMware WWN, the new host port identifier for the addition of the host can let it scan the disk, if not added, but as prompted to wait, So at this time back to the md3200i Add host interface, click on the above by selecting the known non-associated host port identifier to add, through the drop-down menu will see the new value, compared to the VMware iSCSI The value of the Vmbaxx property under Softwareadapter is consistent. Then finish adding the host.

Back to the VMware interface. Scan the disk, will appear in the md3200i definition of the 2 virtual disk, the two disks must be divided into two disk, must not be able to combine the two disks into one, in order to manually redundant, otherwise in md3200i Why do two virtual disk it.

Some time later, you will see two disks made in memory.

650) this.width=650; "title=" 8.PNG "alt=" wkiol1w5e-jbvp5_aafhizq2zpi001.jpg "src=" http://s3.51cto.com/wyfs02/M01/ 70/81/wkiol1w5e-jbvp5_aafhizq2zpi001.jpg "/>c0 and C1 respectively are controllers 0 and 1, remote iSCSI storage is generally to turn on storage I/O Control, click on disk, properties, Select Enabled, the high-level congestion threshold is good by default, unless the network response is slow and then modified.

650) this.width=650; "title=" 9.PNG "alt=" wkiol1w5fmldhytjaag-3omejco247.jpg "src=" http://s3.51cto.com/wyfs02/M02/ 70/81/wkiol1w5fmldhytjaag-3omejco247.jpg "/>

This makes it possible to use the networked storage. This is not done, however, so the default configuration is not load balanced.

Click Manage path in the lower right corner, a new window appears

650) this.width=650; "title=" 10.PNG "alt=" wkiom1w5fawy7oazaalilpwlp-0999.jpg "src=" http://s3.51cto.com/wyfs02/M00/ 70/85/wkiom1w5fawy7oazaalilpwlp-0999.jpg "/>

I this figure is has been modified, the path selection default fixed, this time should see only one activity (I/O), when all 4 are active when the change path selection as a loop, is called polling round-robin, after the modification can see a more activity (I/O), This is the two network card load balancing communication.

A detailed picture can be seen in the screenshot below.

650) this.width=650; "title=" 11.PNG "alt=" wkiol1w5f8khlg8iaafffzekmno537.jpg "src=" http://s3.51cto.com/wyfs02/M02/ 70/82/wkiol1w5f8khlg8iaafffzekmno537.jpg "/> carefully analyzed below.

Esxi03vmnic1 network Interface 192.168.100.31/26 vlan200, When the IP that is discovered by iSCSI and the same network segment has two IP192.168.100.10 and 192.168.100.11, IP is reachable, but the two network cards are divided into two controllers, because md3200i is a force disk binding controller, then md3200i only in IP192 The. 168.100.10 responds to I/O communication, although 192.168.100.11 can communicate in VMware (that is, to find out about dead and alive communication), but not data, another activity of the i/ O is on the 192.168.100.70, this IP is also on the controller 0, see (a) There is a list of IP and controller. Only if the controller 0 is dead will it be transferred 192.168.100.11 and 192.168.100.71, when a network cable is broken, there will be only one network card communication, and the virtual disk binding will not occur from the 0 controller migrated to the 1 controller, the situation becomes the dual network card communication.

From the configuration of the switch, it can be seen that the controller 0 network cable and then two switches. Md3200i's white paper (see page 29th of Powervault-md3200i_setup GUIDE_ZH-CN), which describes a controller's 4 network cards, 22 are connected to two different switches. This continuation method is considered for a long while. Here is the book.

650) this.width=650; "Width=" 971 "height=" 804 "title=" 12.PNG "style=" width:722px;height:770px; "alt=" Wkiom1w5pwpygco1aarsghgpgfi219.jpg "src=" http://s3.51cto.com/wyfs02/M00/70/89/ Wkiom1w5pwpygco1aarsghgpgfi219.jpg "/>650" this.width=650; "title=" 13.PNG "alt=" Wkiom1w5p8wspa9zaamfkqhw1n4727.jpg "src=" http://s3.51cto.com/wyfs02/M02/70/89/ Wkiom1w5p8wspa9zaamfkqhw1n4727.jpg "/>

Originally considered a scheme, that is, all the interfaces of the control are centrally connected to a switch. Server in order to redundancy, the dual network card must be connected to two switches, so that two switches must be able to communicate, to a switch on the network card, and the active controller on the other switch, you need to pass the trunk to another switch, inefficient, occupy the bandwidth of the trunk In addition, if a switch is dead, it is inevitable that the virtual disk will be migrated from one controller to another, which is not conducive to stability, so the migration will not be replaced by bandwidth increase, because there must be a network card because the switch is dead and useless, meaningless. Discard the scenario.

While the Controller interface 22 continues to two switches, then one switch is dead, another switch can continue the original communication controller data transmission, so that the virtual disk will not occur controller migration. Two network cards are continued to two switches, and the traffic is forwarded on two switches respectively, which disperses the flow pressure, otherwise it is concentrated on a switch.

The following is a continuation method of VMware with only two NICs.

In this example, VMware uses vsphere distributed switch in Vcenter, and two network cards are added to vsphere distributed switch for redundancy, and for a machine with only two NICs, use the VLAN 202 name SAN2, VLAN 203 name SAN3 as a storage segment, this host's network card is also connected to the two switches, has done vlan202 and 203, and in Switchport trunk allowed VLAN added 202 and 203

Interface GIGABITETHERNET0/15
Description Esxi01vmnic0
Switchport Trunk Encapsulation dot1q
Switchport trunk allowed VLAN 5-12,100,101,202,203
Switchport mode Trunk
Storm-control Broadcast level 10.00
Storm-control Multicast Level 5.00
Storm-control Action shutdown
Spanning-tree Portfast Trunk

Enables VMware's distributed switches to access the stored VLAN segments.

The following are the distributed port groups to which VMware's distributed switches can be mapped.

Open the VMware vcenter Managed home page-"network-" to create a new port group,

650) this.width=650; "title=" 14.PNG "alt=" wkiom1w5rvlx8gauaacua2qenfa413.jpg "src=" http://s3.51cto.com/wyfs02/M00/ 70/89/wkiom1w5rvlx8gauaacua2qenfa413.jpg "/>

Create two, one vlan202, one vlan203.

In the host-"configuration-" network, see the vsphere Standard switch content is empty, click Open vsphere Distributed switch, click on the management virtual adapter, there should be at least one vmk*.

Add, new virtual adapter, only Vmkernel optional, Next, select the Port Group drop-down menu to find the dvSAN2 you just made

650) this.width=650; "title=" 15.PNG "alt=" wkiol1w5slqatkbeaaeekh6l8ea471.jpg "src=" http://s3.51cto.com/wyfs02/M01/ 70/86/wkiol1w5slqatkbeaaeekh6l8ea471.jpg "/>

Also here can not choose vmotion and FT, also can not be used to manage traffic, with IP address: 192.168.100.141/26

Continue to do a VMK, port group selected vlan203, with IP address: 192.168.100.211/26

And then just like the exsi03 machine in the storage adapter, with the dynamic discovery 192.168.100.141, complete the Add, do not forget in the md3200i also add Esxi01 host Oh.

Because Vcenter is in the global management, this time in the memory directly can see the esxi03 on the two virtual disks, do not configure, automatically appear.

Although esxi03 is accessed with IP192.168.100.10 and 70, ESXI01 uses 192.168.100.130 and 200 access, but vcenter automatically identifies them as one. What is the reason that can store content in traffic, See a. naa.6c81f66000ccc02500000960555464ff such a folder, inside there is a slotsfile file,. naa.6c81f66000ccc02500000960555464ff is actually the stored virtual disk global identifier: The value behind 6C:81:F6:60:00:CC:C0:25:00:00:09:60:55:54:64:FF,.NAA is the same as the virtual disk global identifier in the md3200i. Depending on this determination is the same disk, although the access path is different.

This article from "Genius without that 1% is absolutely impossible" blog, please be sure to keep this source http://xushen.blog.51cto.com/1673219/1679897

VMware vSphere5.5 Connection md3200i details (ii)