LUN 20 TO - CX3 / PS5000

Similar Questions

• Change of strategy by default multichannel in ESXi 5.0 has no effect to the existing LUN

  Hello.

  HISTORY:

  We have several hosts with ESXi 5.0 connected (via zFCP) to SAN (a Clariion CX3 and CX4 a) 2.

  After installing ESXi (a few months ago), we've changed the strategy of multiple paths by DEFAULT to RoundRobin for both CLARiiON systems, but NOW we wan to return the default policy for the CX3 on MRU leaving the default policy for the CX4 as RoundRobin.

  * Cause CX3 is not compatible with ALUA but CX4 is, we simply changed to MRU the DEFAULT policy for the Type of array VMW_SATP_CX storage only.

  We executed the following command:

  esxcli storage set - default-psp VMW_PSP_RR - ATAS VMW_SATP_CX ATAS nmp

  After that, we restarted the ESXi to "activate" the changes.

  THE PROBLEM:

  After the reboot, surprisingly existing LUNS of the CX3 continue to use the RoundRobin policy, even if the Type of storage array appears in the form VMW_SATP_CX and the command 'esxcli storage nmp ATAS list' shows that VMW_SATP_CX now use MRU as the default policy.

  I manually changed the policy for some LUNS of RoundRobin on MRU and then restarted the ESXi and change persists for those LUNS (using MRU rather RoundRobin) but the rest of the LUNS continue to use RoundRobin.

  Also, I created a * NEW * LUN in the CX3 and the ESXi it displays correctly (using MRU rather RoundRobin).

  THE QUESTION:

  So... to change the DEFAULT policy with esxcli is useless for the existing LUN and I have to manually change EACH CX3 LUN to SRM strategy to complete the process?

  .. .or maybe I forgot any action a step further?

  Kind regards.

  THE QUESTION:

  So... to change the DEFAULT policy with esxcli is useless for the existing LUN and I have to manually change EACH CX3 LUN to SRM strategy to complete the process?

  .. .or maybe I forgot any action a step further?

  It's true... Change the default PSP will affect any new LUNS added.  Your existing LUN needs to be changed

• RAW Device Mapping / beginner question

  Hello list,

  I'm totally new to RDMS, so I have a basic question. I have a snapshot of LUN (Dell Clariion CX3-20) to a NTFS logical unit number (1 TB, a lot of Windows file as doc, pdf, etc.).

  Can I mount this snapshot in a virtual machine? Can I use RDM to achieve?

  If so, is - anyone can explain this in detail a bit?

  Kind regards

  Andreas

  The question was:

  I have a snapshot of LUN (Dell Clariion CX3-20) to a NTFS logical unit number (1 TB, a lot of Windows file as doc, pdf, etc.).

  If a snapshot of a physical system is the answer is Yes (wink CX are the same as LUNS).

  If it is a snapshot of a computer virtual with a RDM, the answer is "maybe" (you must play with LVM advanced options).

  André

  * If you found this device or any other answer useful please consider awarding points for correct or helpful answers

• EqualLogic PS5000 - how many LUNS

  We just bought a Dell Equallogic PS5000 to connect to our virtual machine environment.  We will connect only ESX host (3 of them) to the Equallogic.  The guest VM will use local storage to ESX, but all virtual machines running on the hosts will get their memory by the Equallogic.

  The question is: how many LUNs we set up on the PS5000?  What determines how many LUNS are necessary?

  Thanks for any information.

  "tuck,

  > We need to have an another LUN for snapshots of server (equallogic)?

  N ° when you create a volume on your Equallogic SAN, it will default again reserve the same amount of space for the snapshots. At any time, you can increase or decrease this placeholder. If you use this placeholder the default action is to delete the oldest snapshot.

  > Can initially set you 1 LUN with a limited amount of space (say 1.5 to) then expand out more later if necessary and add additional LUNS required?

  Yes, but you can't do that. Yes, you can create a volume of 1.5 TB and later it increases the size, but it is the new size will be automatically recognized by the operating system. I do not think hand VMFS (VMware f system Island) will stretch to see increased at all space, so your option here is to create a second volume with the point and creating a large volume. The downside of this is that something happened to one of the targets in a store of data scopes ALL data are lost. I agree that extended are bad - they placed there, and there, they must be kept.

  However, lets say, you know that in the future a specific target (LUN) must be bigger, but because you have only 7 TB of space you can not afford to provide it is all he needs. For example a fileshare, you know over time people are just going to store more and more, it's where you could introduce a large volume, say 2 TB, but it would only consume as much space as the data required stored summer, as more data is written more space is necessary. This is called a provisioning, this is a double-edged sword. If you have too use a provisioning by presenting each volume as finely put into service and so this 70 TB of storage of your 7TB dashboard, what you can do, but if you neglect to submit buy a req to buy and additional storage beyond the gates 7TB before starting the people write about it, you will cause yourself some pain as the operating systems do not like been said they can not write no more!

  Do take a step back and figure out your available LUN, but have no more thing, with Storage VMotion, you can always move VMs. I would advise against the creation of a LARGE volume and stick all your virtual machines on it, rather I would stick with volumes ~ 500GB and spread them out. Don't let some soft space by volume for the use of snapshots of the virtual computer.

  > Are affected LUNS based on some hard disks, storage or both?

  It is not clear to me what is your question here, please provide details, but to confuse the Equallogic issue tables do not have LUNs, but presents goals, but I think that your point of view it is not a difference.

  You may want to get a copy of http://www.amazon.com/VMware-Infrastructure-Advanced-Technical-Operations/dp/0971151083

  Regarding the RAID configuration, if you feel that you are wrong, you can always change it without data loss, you can set per LUN or let it default to what you choose for the entire table.

  Good luck.

• CX3-20: host in several storage groups?

  Hi all

  What is the trick? I have all the storage groups created on our SAN access to our server VCB (Virtual Consolidated Backup of VMware).

  The problem is that our Clarion EMC Cx3-20 allow us only to add a host to a storage group.

  If I want to connect the VCB server to a different storage group, I get a prompt saying (in Navisphere):
  AMS - back.corp.com: this host is already connected to a different storage group. Connection to the current storage group will be be disconnected from the other storage group. This will cause the host unable to access the LUNS that are in the other storage group.

  VCB server must see all LUNS, can advise you?
  Thank you

  
  

• Como can present LUN is a a physical server

  Tengo una SAN Clariion CX3 - 20FC in the LUN creamos cual y los a los servidores office investigation...

  Ahora lo q tengo q hacer are present LUNS to a servidores unos Fisicos pero no las he found como hacer para LUN present is an a physical server

  Saludos thanks for the help

  Dear Warrenjc

  Amendments ordered dar mas datos.

  Quieres simplice acceso a las same LUN as los esx o an otras LUN?

  Saludos

• getting dropped under heavy load of iSCSI LUNS

  Hello

  
  I found a few forums on this... but not a response yet.

  I have a Hp Lefthand 4300 cluster for the storage, connected to 2 x DL360 G7 ESXi 4.1 via iSCSI (Hp edition) guests and a Procurve gigabit switch dedicated for iSCSI. Whenever I run a backup down the LUN for a few seconds (usually 5, however it was as much a minute).

  Things, I tried...

  Updated software from the left to the latest version

  Updated all the HP firmware on the hosts

  Update to the latest drivers HP Boradcom on ESXi hosts

  Changed the timeout 1400 (esxcfg-advcfg - s VMFS3/14000/HBTokenTimeout)

  These servers come into production in the new year... Please please can someone help me solve this problem?

  I call open with Vmware and HP, nor was particually useful, Vmware has suggested I called HP and updated card drivers NETWORK, HP simply said that it was a network problem and closed the call... Thank you!

  I intalled surveillance under tension and can not see all ports dropped etc. All the equipment is new and the LAN cables are Cat6.

  Don't really know where to go from here!

  The exact error I get is...

  Lost access to the volume (4cf7b1ac-5fdcc218-5c2e-1cc1dee48e30)
  S1 - SP1OS) because of connectivity issues. Recovery attempt is in
  progress and results will be announced soon.
  Info
  2010-12-27 01:06:06
  192.168.32.2

  Please help me!

  Kind regards

  Kris

  We have recently changed some hosts who also had problems on a heavy load with iscsi San connections. We used hwiscsi (TOE), Broadcom adapters (was a recommendation from different vendors) before and this has recently changed to swiscsi (with Jumbo).

  No. dropped / resets on connections over this host. And a much faster flow of vm (tested with IOmeter).

  We also saw a much beter performance (throughput) on hosts with NIC configured to swiscsi on comparisation of hosts with an iSCSI HBAS (Qlogic 4062).

  We use VMware ESXi 4.1 (latest patches).

  Problems with broadcom TOE (hwiscsi) on HP DL360G6 also on Dell M610 blades.

  SAN: EqualLogic PS5000 PS6000, PS5500, PS6500,

  Switches: Juniper weer Jumbo Frames and flow control.

  So I recommend you also use swiscsi on these broadcom adapters.

  If you are looking for broadcom issues with vmware, find others with similar experiences.

• PS5000 & PS6000

  Hi guys,.

  I have a scenario where I have a cluster of ESX 4 with a SAN PS5000 as back-end storage. We are looking to buy a PS6000 and this area in the same cluster, and then migrate the virtual machines to it.

  If I presented the two without via iSCSI is there problems with both SAN and with the creation of LUN on the new San with the same LUN ID etc?

  Thank you

  DC

  Well, then you can:

  Create a new group of EqualLogic and installation your LUNS properly.  Then you can add the new group as a target iSCSI on all of your guests and add the LUNS as VMFS targets.   Once you have the new LUNS, you can Storage vMotion virtual machines you mention.

  Once you are convinced that everything is off the PS5000 (check, check and triple-check), then delete all VMFS warehouses based on the PS5000.  Then, you need to remove the IP address of the PS5000 storage group in the list of targets iSCSI on each host.

  After that, I recommend you restart each host and let to clear the list of targets of the old LUNS automatically, otherwise, you will need to manually remove all targets in each host.

  -

• EMC Clariion CX3-80

  Hello

  We'll put our vSphere on EMC CX3-80 with 600GB 15K drives RAID5. Our plan is to have about 30 discs so EMC DAE 2.

  Which were not sure is the best way to configure the LUNS and RAID for optimal performanceeg. Use the metaLUN, or simply through the same DAE?

  Also what size the VMFS should be and whether it would be a good idea to use VMFS extensions. Typically, we will have virtual machines where the OS drive is on 70GBs and the D:\ can vary from 200 GB to a few larger 400-600GBs.

  Thank you very much

  The reason that I thought holding through the DAE 2 metaLUN using the 600 GB drives to ensure that we get allows to optimize more SPs rather have LUN through a single DAE using 1 SP.

  This could be a good reason.

  But you can quite the same reach by using multiple LUNS owen by different SP.

  We'll probably use RAID5 with 4-5disks by RAID group and a global spare wheel.

  If you have multiple disks also look to build a group of Raid RAID10.

  But what about the size of the block VMFS - y at - there no performance impact?

  None.

  But the alignment could help. See: disk partition alignment

  In our case, we will use very probably the last 800 GB but it also means my VMFS must be mapped to each LUN 1VMFS thus: 1LUN.

  800 GB IMHO is a good size and is better to have 1:1 for VMFS/LUN and Raid group.

  But the VMFS extensions should be used

  I'd rather not use extensions.

  And also if some VMs have d:\ e:\ f:\ which can go up to more than 800 GB, which would be the best approach

  Consider using the RDM disk in this case.

  André

• VMs per LUN

  Good evening

  <! [endif] > <! [if gte mso 9] >

  Today, I had

  a specialist of the virtual machine in my company telling me I should have 1 VM per LUN or

  Otherwise, the performance would be poor. I told him that it was impossible and the

  Another suggestion was to have all the VMS each running on an ESX host LUN. I ask

  him how I could do that without jeopardizing the DRS and I didn't get any answer.

  1^st

  -Is it true?

  <! [endif] > <! [if gte mso 9] > 2^nd

  -What to do

  best practices in question are to know that my company has no HARD drive

  space for waste and my storage is basic fiber (CX3-10 DELL)?

  Filipe Pinto, D. C.

  I want to say can I consolidate all virtual machines of a lun to a single host?

  If you are using a shared storage and relying on the DRS and HA, then you will have all the shared storage as LUN presented to the whole of your ESX hosts.  He didn't need to be any kind of an ESX host to a LUN relationship.  You might have a host with 5 LUN or share of 5 ESX host with a logic unit number.  The general idea is to split the charges between hosts and LUNS - there are many different ways to do this.

  Is it true that performance degrades if on 1 lun I have has 2 vm, a run on host 1 and host 2? Is that creates any kind of locks?

  Basically, that a single ESX host never would have one virtual machine files open at any time.  Seeing her and access are two different things here.  See p.6 of the Study of scalable storage Performance for information on SCSI reservations and how they are treated.

• Free space on the PS5000 and host different reports

  Hi guys,.

  I got my hands on a PS5000 a few days ago and am still getting used to it.

  My question is perhaps more approprriate for a Dell forum, but I'll take my chances.

  I just emptied on a volume, making use of svmotion. vCenter reports my LUN vacuum

  but when I check the PS5000, he pointed out that there are still some 160 GB in use.

  It seems that other volumes reported sizes are inconsistent with what ESX reports.

  Any ideas? I tried clear as a disconnect, and refreshing

  See you soon

  Where my question on verification of i/o to the LUN. Pulling up to the LUN, while the host is using is not a good idea.

  Property to true.

  In any case, I'm not sure that there is a command (except esxtop) line to monitor the disk i/o.

  André

  * If you found this device or any other answer useful please consider awarding points for correct or helpful answers

• Emc LUNs move to another host esxi

  We have an ESXi server connected to an emc clariion cx3 - 10 c (iscsi, 1 logical unit number). We will fully ESX and this server needs to be rebuilt to that. I have built a new server esxi and want to keep the virtual machine residing on the lun. Is there a way to communicate the number of logical unit on the new server esxi? I see an option migrate in Navisphere, don't know what it does. I am a rookie!

  You should be able to present the existing LUN as your return VMFS datastore to the new ESXi host by adding the new host of the storage group on the CLARiiON. Then the new server will see the VMFS datastore and as long as (Vswitches) network settings are the same on the new host of ESXi, you can migrate virtual machines to the new host and then dismantle the old host.

• SQL 2005 Volume Design on EqualLogic PS5000

  I am in the process on the implementation of a SQL 2005 server in ESX 3.5 utilizating an Bay Dell EqualLogic PS5000 (16x15K SAS drives).  My question is on the volume / lun considerations.  Because the equallogic use virtualized storage does really matter if I put separate log and DB on Lun volumes?  Any lun volume goes on inputs and outputs should be spread across all disks in any case, if I understand correctly.  Is there anything else I should think to when planning on volumes of sql in ESX?

  Any advice would be greatly appreciated.

  You must look in Workgroup Manager to see what connections use port Eth.  Regarding iSCSI, if you use a HW iscsi device you can use several volumes and make the poor load balancing to balance the use of the road.  The Equallogic will do a decent job, balancing connections to ports different Eth.  You can't control which Eql Eth ports are used for a given volume.

  If you use iscsi sw IP VMkernel on the vSwitch will use only 1 path to each port EQL Eth IP address.  So if you have 2 sw iscsi NIC in a vSwitch and several volumes, I hope that you will be on > 1 port Eql Eth and then the IP vmkernel will balance the loads on the 2 natachasery.

  Now that I have more information, I think that more volumes are better in your case.

  You cannot grant a lot about EQL bays, including the use of processor cache or volume.  There is no publicly available documents that describe the inner workings of the tables.  Several volumes is fine, I think the max is 512 or 1024.

  Ben

• Access via SSH o/s RN102 iSCSI LUN

  Hello

  I can access NAS via SFTP & SCP.

  Where in the structure of the file is the LUN file so that I can copy off the coast of the NAS for editing also.

  We have an iSCSI partition that is only considered as RAW by the iSCSI connector in Windows and we need to see if he really is totally damaged,

  The data\LUN file appears to be empty.

  Andy

  It's in a hidden directory .iscsi

• 'Convert' part in LUN for iSCSI access

  Hi all

  I used my RN2120 as a SMB share for my exclusive Microserver HP for the content to Plex. I have now come to install a double extra NIC in my server and I would like to use the ReadyNAS as iSCSI storage.

  What I can't work, is to "convert" my SMB on a LUN sharing for access via iSCSI.

  Someone at - it an idea? Or should I return my data (~ 5TB) on my server, delete the share, create the LUN and bring back the data? #fingerscrosseditsnotthissolution

  Thank you very much

  Carl...

  There is for example a NTFS file system THAT put on the iSCSI LUN

  You will need to copy the data to your PC, delete off the coast of the NAS, create the iSCSI target and copy the data back.

  If all the data you have, then after copying data off isn't 5 TB of data you can update the firmware, verify that the update of the firmware successfully, do a factory reset, create the iSCSI target, install it on your PC using the initiator and restore your backup data.