pRDM vs vRDM

Similar Questions

• interchangeability virtual / physical vRDM vs pRDM

  Hello

  I am currently implementing a replication solution in our vSphere environment 5 and would like to be able to also reproduce some virtual machines that are currently using pRDMs. Because snapshots are a requirement, I know I need to change the pRDMs corresponding to vRDMs as described in http://kb.vmware.com/kb/1006599.

  A reason for us to use pRDMs was the interchangeability with the physical host computers. In other words, if the virtual machine goes down we have yet to introduce the LUN to another host (physical). And Scott Lowe's book "Mastering VMware vSphere 5" States the following about the difference between pRDMs and vRDMs:

  Another important use of pRDMs case is that they can be presented in a virtual machine from one physical host regardless. This gives pRDMs a flexibility which is not found in RDMs virtual mode or virtual disks.

  However, I've adjusted tested interchangeability with a vRDM and discovered that I can introduce a vRDM containing an NTFS partition to a physical host with no problems. That, in theory logical I suppose, because the two types of RDM are - as far as I understand - a few pointers to LUNs, with the difference that pRDMs submit all SCSI (with the exception of the REPORT LUNS command) orders on the device and vRDMs presents only the commands READ and WRITE on the device.

  Statement of Scotts on vRDM interchangeability is just a small flaw in his book, or did I miss something?

  Thank you

  Michael

  Hi Mike,.

  You're right, it's a small flaw.  I worked with RDM for the last year and a half, and I've always been able to present vRDMs to the physical host computers.

  Kind regards

  Feddaoui

• Conversion of pRDM in VMDK

  I would like to convert a virtual machine using pRDM disk VMDK disks, so that I can backup using Veeam.

  I plan to stop the virtual machine and then migrate to a different data store to convert the pRDM in VMDK. I did this on a test computer and it seems to have worked OK. However, I have noticed on the test machine the VMDK which had been converted from pRDM, was the possibility of "independent - persistent ' selected on them.

  I guess that's because the pRDM is configured this way by default?

  Can I remove the independent - persistent option without negative effects for the virtual machine?

  
  Thank you

  Fix. RDM disks are unaffected by snapshots, if after the migration of VMDK also inherits these parameters.

  Be sure to uncheck this. Otherwise in certain circumstances you can have problems with backup VM as well.

• 25 to drive...  VMDK or pRDM?

  I have a client who wants a single large 25 to disk attached to a virtual machine. The ESXi server is 5.5U1, and everything is compatible for both.  pRDM or a TB 25 VMDK, which one should I choose?  This will be the data reader for a repository of content in many files.  I think because of that, I want to can I use a client backup rather than backups VADP.  But who should I choose for the LUN type?  EMC VMAX for storage.

  I also would support a virtual disk, that better.

  cases of use of physical RDM Mode and limitations are listed below,

  http://www.vmwarearena.com/2013/02/raw-device-mapping-RDM-use-cases.html

• Different VPD on guests when using pRDM

  Hello

  Repost this in the correct area and hope some thoughts.

  In the process of creating a cluster across boxes using two hosts the SQL. The SAN is Dell MD3200i with multi-unit MD1200 expansion. Has created several new LUNS which manifest the same LUN id when shared between two hosts. Using the VMWare best practices. The LUNS added to guests as pRDMs and appear very well. Can I use the disks of each guest taking offline as required and tested with basic files. All normal behavior. However, the list posted of MSCS disks is showing VPD page 80 h as different a guest at the next serial number. Page 83: UUID is identical, but the difference in serial number causes MSCS see the disks as differing and therefore cannot be used as cluster disks. The disks have been added using new SCSI controller with the same SCSI id through the two guests. I deleted the LUNS, controllers SCSI, maps several times now, but the series is always different. With the help of Windows Server R2 2012 for the guest and VSphere 5.5 Update 2 for guests. I'm really bad with that, will try to recreate the guests but can not understand how the same LUN can report a different serial number. Still using LSI logic SAS. Is it possible to query the VPD 80 h of vCLI page to determine what comments does not? It is more the fault of the SCSI controller?  Any help would be greatly appreciated.

  Thank you

  Jonathan

  Quick update - we have updated our 3rd host to ESXi 5.5 Update 2 and fixed a minor issue of memory on host 2. Has also had a PCIe bus failure that has been resolved by reinstalling the risers. After all this, the CAE now reports the correct serial number. Still no idea what was causing the problem but thankful it is now resolved. Maybe the multiple reboots and the PCIe bus failure could be a cause.

  Thank you

  Jonathan

• Need to convert vmdk using pRDM vmkfstools command

  Need to convert to pRDM vmdk format command. tried to convert using cold migration that it stuck at 23 percent. Are there vmkfstools or powercli to convert?

  You can try the host command line is:

  vmkfstools -i /vmfs/volumes/source-ds/source-vm/source.vmdk/vmfs/volumes/target-ds/target-vm/target. VMDK

  You can optionally add thin d in case you want to that thin target to be configured.

  What you need to do - before running the command - is to ensure that the size of the disk in the hard file corresponds to the actual size of the LUN in case the LUN has been resized / increased at some point.

  André

• pRDM to pRDM migration?

  I use a lot of pRDMs in my environment vsphere (v 5.0 U2), because we have a lot of SQL Server, MS Clusters. However, it is time to buy a new SAN, and I wonder what is the best way to go about it. I remember an old article by Cormac Hogan and one of the questions that he had been, how useful would a RDM to RDM migration tool? At the time, I didn't have on this subject, but now that I have, it would be very useful. Not sure if this feature is available now, though.

  I know a solution would be to turn VMDK, but since I use MSCS, which is not possible.

  I'll upgrade to vSphere 5.5 in a few weeks, do not know if this changes anything, regarding this question

  Thank you

  Mike

  I'm not aware of a pRDM migration tool and again, you will have to do it manually from the command line. Anyway, most major storage vendors offer tools SAN Copy (usually free) that can even migrate LUNS in other storage vendors. It is perhaps the best way to migrate from the IMO.

  André

• VM VMware Data Recovery and save after pRDM vdisks

  Hello

  We will use VMware Data Recovery. A virtual machine will be some vdisks that will be mapped and pRDM. I know that vSphere can't save pRDM. What I would like to know is if VM with pRDM can be saved by the device of data recovery.

  So, can I add VM who pRDM unit of data recovery. Vdisks (and associated data) that are no pRDM don't rely on VMware Data Recovery?

  Thank you!

  Hello.

  See KB 1015488.

  Good luck!

• Serial number different VPD when using MSCS for pRDM

  Hello

  I'm trying to set up a cluster across boxes using VSphere 5.5 SQL update 2 and R2 Windows Server 2012. Following supported and recommended configuration, but met with a blocker. The RDM are presented to the two guests (power off). When online, the disks are configured and can be used by both hosts respectively. No apparent problems. However, when run the failover cluster validation, CAE serial number is different on the guests. Page 83: UUID is the same but the serial number is different. I think that it is the information on page 80 h. This can be changed? Why is it different for the same LUN presented as shared storage. The logical unit number ID is the same on each host. Failover cluster considers different disks as a result. Any help would be greatly appreciated.

  Thank you

  Jonathan

  Quick update - we have updated our 3rd host to ESXi 5.5 Update 2 and fixed a minor issue of memory on host 2. Has also had a PCIe bus failure that has been resolved by reinstalling the risers. After all this, the CAE now reports the correct serial number. Still no idea what was causing the problem but thankful it is now resolved. Maybe the multiple reboots and the PCIe bus failure could be a cause.

  Thank you

  Jonathan

• convert RDM VRDM

  Hi guyz,

  I'm not used to working with raw device mappings, but I have a little problem somewhere.

  I searched the forums, but I've still found all the answers.

  I need to convert a raw device mapping to a virtual raw device mapping.

  The reason for which it has been used a RDM is not clear to me, the database running on this partition is only the larg of 20 GB or something like that.

  It is necessary for reasons of backup (veeam)

  You can select the physical mode or virtual RDM when you create the RDM.

  Stop the virtual machine, remove the ROW of your config and add the LUNS, and this time,.

  use the virtual rather than physical.

  -KjB

• All LUNS assigned to ESXs

  Hello
  I want to know some information about a virtual infrastructure. For example:

  "ESX host", 'LUN Id', 'Capacities', 'Type '.
  'ESX1', '12', '400 GB","flat ".
  'ESX2', '12', '400 GB","flat ".
  "ESX1", "15", "250 GB", "flat".
  "ESX2", "15", "250 GB", "flat".
  "ESX1", "25", "100 GB", "vRDM.
  "ESX1", "33", "200 GB", "pRDM.

  Or even better for me:

  "ESX host", "LUN Id", "Ability", "Type", "Datastore?
  'ESX1', '12', '400 GB', 'flat', 'DataStore_1 '.
  'ESX2', '12', '400 GB', 'flat', 'DataStore_1 '.
  "ESX1", "15", "250 GB', 'flat', 'DataStore_2 '.
  "ESX2", "15", "250 GB', 'flat', 'DataStore_2 '.
  "ESX1", "25", "100 GB", "vRDM', 'n/a '.
  "ESX1", "33", "200 GB", "pRDM', 'n/a '.

  
  Is - this popssible?

  Thank you

  Best regards.

  OK, I've updated my script with a couple of protections.

  And I threw in the Monday.

  Have a go with the attached script.

• Convert physical RDM to RDM virtual disc Exchange

  Hi all

  I need to convert a shared disk (SAN LUNS) used by Exchange 2007 SP1 of RDM physical to virtual RDM, I would like to have a confirmation fo this strategy:

  1. Turning off the virtual Exchange Server computer

  2. Remove the disc in the virtual machine configuration vRDM mode (not the disc, of course)

  3. Fix the disc in mode pRDM

  4. Turn on the virtual machine

  Question:

  This stratey is correct?

  could I have a problem with the disc will type?

  Thanks in advance

  Angelo

  Hello.

  That's it – you can document the file mapping according to how many of them you have.  The procedure is also documented in KB 1006599.

  Good luck!

• Errors SCSI sense in Event Viewer

  Once an hour, I get a handful of warning message in my event viewer.  They are all the same except change that denotes a different physical disk (7-14).  There is a mix of SATA and SAS in a MD1000.  The SATA drives did not produce the error.  A so-called failure last week SAS drive after having been online for only 1 week.  All green lights on the MD1000 and OMSA shows everything is normal, but I get the warnings in the event log.  Can someone explain what the error or the CDB is the cause?  Should I worry about this?  This is causing readers to fail?

  Data sheet:

  PE 2900

  PERC 6/e w / 512 mb

  MD1000 7 Seagate 1 TB SATA disks and drives hard Seagate 8 450 GB SAS.

  The message in Event Viewer:

  Event type: Information
  Event source: server administrator
  Event category: Storage Service
  Event ID: 2095
  Date: 02/09/2008
  Time: 09:49:13
  User: n/a
  Computer:
  Description:
  Sense key sense data SCSI: 5 detection code: detection 24 qualifier: 0: physical disk 0:0:7 controller 1, slot 0

  An excerpt from the newspapers of OMSA Perc6 controller:

  02-09-08 09:49:03: EVT 08954-09/02/08 9:49:03: 113 = feel unexpected: PD 12 (e0x11/s11), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a0752800) Cmd = 3, Sns = a0752840 CIH [12] Tgt 10

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08955-09/02/08 9:49:03: 113 = feel unexpected: PD 1a (e0x11/s14), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a062a800) Cmd = 3, Sns = a062a840 CIH [1A] Tgt d

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08956-09/02/08 9:49:03: 113 = feel unexpected: PD 1 b(e0x11/s13), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a0781800) Cmd = 3, Sns a0781840 Tgt DevId [1B] = e

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08957-09/02/08 9:49:03: 113 = feel unexpected: PD 1c(e0x11/s12), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a07dbc00) Cmd = 3, Sns = a07dbc40 CIH [1 c] Tgt f

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08958-09/02/08 9:49:03: 113 = feel unexpected: PD 1e(e0x11/s10), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a05f9c00) Cmd = 3, Sns [1] a05f9c40 CIH = Tgt 11

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08959-09/02/08 9:49:03: 113 = feel unexpected: PD 1f(e0x11/s9), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a05b2e00) Cmd = 3, Sns = a05b2e40 CIH [1f] Tgt 12

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:03: EVT 08960-09/02/08 9:49:03: 113 = feel unexpected: PD 20 (e0x11/s8), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:03: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:03: REQ ILLEGAL: pRdm (a05cf400) Cmd = 3, Sns = a05cf440 CIH [20] Tgt 13

  02-09-08 09:49:03: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:03: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  02-09-08 09:49:04: EVT 08961-09/02/08 9:49:04: 113 = feel unexpected: PD 21 (e0x11/s7), PEH: 12 01 00 01 1 dc, meaning: 70 00 05 00 00 00 00 0 a 00 00 00 00 24 00 01 00 03 c8

  02-09-08 09:49:04: DEV_REC:IllegalReq - CBD: 12 1 1 1 d 0 ms

  02-09-08 09:49:04: REQ ILLEGAL: pRdm (a0522a00) Cmd = 3, Sns = a0522a40 CIH [21] Tgt 14

  02-09-08 09:49:04: info: aslen 0, = a, cmdSpecific [3] = 0 [2] = 0 [1] = 0 [0] = 0

  02-09-08 09:49:04: asc = 24 Ascq 0, Sks:c8, fp [0] = 0, fp [1] = 3

  -Jake

  
  

• Limitations of architecture with replication of VM with physical RDM Mode

  What are the Limitations of Architecture with replication of VM connected with physical RDM Mode in vsphere replication. Why VMware does not support this?

  I will add some colors to the response of GS. A particular interest with regard to the physical path RDM (pRDM) works, summarize these two chips of the article:

  • Physical mode specifies minimal virtualization SCSI of the mapped device, allowing greater flexibility for SAN management software.
  • VMkernel passes all SCSI commands to the device, with one exception - the REPORT LUNS command is virtualized, so that the VMkernel can isolate the LUN to the virtual machine owner. Otherwise, all the physical characteristics of the underlying hardware are exposed.

  There is an integrated in vSphere vSphere (RV) replication agent that has several functions. In particular, she keeps track of the Scriptures to the virtual machine. When a replication cycle occurs, the changed data is replicated to the target location. The VR agent should be able to 'see' These Scriptures to follow. Given that the physical mode all SCSI commands directly to the device, the VR agent is unable to follow these changes for replication.

• VM shows physical mode as unreadable RDMs after DRS put VM on new host

  Need help with a problem:

  a virtual machine with pRDMs which is part of a VCS configuration has been closed, host migrated manually again, then power on, DRS he moved to yet another host.   and for some reason, the RDM are now unreadable

  Does anyone have an idea why this happened and how to fix this?

  Notes - for this virtual machine DRS has been on 'manual', while the clusters more DRS setting is fully automatic.     Because it is set to manual, I wasn't the one that vaulted to the top of the virtual machine, but I guess the user who did it was invited to place the virtual computer somewhere and he chose to place it on a new host then that it had migrated to.

  Made this recommendation of the DRS and move them to the new host clustered VM ability to see the pRDMs?

  In the guest OS I run diskpart and the disks are online, but appear as unreadable.

  Thanks for advice on this one to help us understand the behaviour here, and what could have caused this.

  I am perfectly aware that the VNMware cannot vmotion VMs with shared during an attempt of vmotion SCSI controllers.  But it's not clear to me that the process is under the covers when DRS wants to move a virtual machine somewhere in the power.

  DRS uses vmotion to move VMS between hosts and it should has failed, if they are configured for shared scsi - I do the bet when the virtual machine has been powered and he was placed on the host wroong who has no access to the RDM - powering the virtual computer if you start on the correct host do you have access ot the RDM?