SRM Test AD/DNS vs AD/DNS recovery failover recovery

Similar Questions

• Issue of SRM Test Recovery plan

  Hello

  We have two sites with hosts 5.5 ESXi and vCenter 6 and 6 SRM installed and configured on both sites. We use IBM midsize v7000 storage on both sites. I configured SRM with array based replication. Everything seems fine, but when I run the plan of recovery in recovery of test mode, it generates the error on the stem 4;

  --------------

  Cannot create snapshots of the replica devices.

  Cause:

  Failed to create the snapshot of the consistency group replicates Test.

  Cause:

  SRA command 'testFailoverStart' failed for the consistency group 'Test '.

  Cause:

  Replication may have ended because of a previous failover operation.

  Refer to IBM SAN Volume Controller troubleshooting

  ------------

  Home is also the snapshot. Please, no matter what advise how to solve this problem.

  Thanks in advance...

  Try to delete the file THAT SRA-STATUS located in C:\Program VMware vCenter Site Recovery Manager\storage\sra\IBMSVC and then refresh/rescan your picture frames.

• DR SRM Test: activate Resingature

  Hello

  We have SRM 5.1, 5.1 VC and ESX 4.1 in our environment.  I'll perform the test of DR. Do I have to value the parameter enable re-signature 1 manually on the site of recovery, before running the test from DR.

  Amit

  I raised the SR for it and got the following response from VMware

  LVM. EnableResignature must be set to 0 before the start of the trial of DR failover.  SRM set to 1 during the recovery plan and after that completion of the LUN must be automatically reset to 0.

  A

• SRM Test Layer2

  We are preparing for a trial of DR in April and run a couple of tests "minis". We run cx480 EMC with Recoverpoint. We bought SRM and with the Layer2 up works perfectly. But during our test we similate the disaster by 'cut' the Layer2. But when we run a test of our stimulus package, it fails with the error - failed to create snapshots of the replica devices. Failed to create the snapshot of the replica 12044 consistency group. SRA command 'testFailoverStart' failed for the '12044' consistency group. My understanding is that it works with a recovery instead of a test but I am quite nervous about recovery once our scenario is completed. I don't want to back changes made during the trial of return to our Production data center

  Any help or suggestions are appreciated.

  If your recovery test works, then as long as your LUN is mirrored without error, this error is 'normal', as it tries to replicate before the failover and the rest of your Recovery Plan should continue successfully. If you want to be sure that your virtual machines will rise on recovery Site properly with the data before disconnecting network SAN or cracking, then do a planned first migration.

• SRM Test

  Hi all

  I just want to ask, in the course of trial SRM (press the button 'Test'), is it true protégé, and site recovery must have interconnection between them? I ask because I failed a trial SRM (partial DR) recently and one of the requirement during this partial DR is to break the interconnection between HQ & DR site, it make sense the error I got was related to SRM connection between protected & recovery site issue. BTW, I have no problem real failover (press the 'Run' button) so far.

  And no problem so run away RS test for other clients since we don't actually break the link between protected & recovery site.

  My configuration:

  • SRM 4.1 + last SRA
  • HP EVA 4400 + 9.03 view command

  Thank you

  Hello

  From what I could find seems that MRS does not require connectivity between sites.

  However some securities regulators is required (at the level of storage). I have no saying which of them need and are not in the list.

  You can contact HP to check if this is necessary for your SRA.

  Michael.

• SRM test - failure storage caused Reset

  Hi all

  Everyone had a problem during the test SRM that when the test finished his last step, so: -.

  11 reset storage Post Test

  They have experienced such a load on their SAN that it disk latencies and cause outages for all virtual machines on those LUNs?

  We have done so far here two tests and have experienced this, when you make more than 4 LUNs, essentially when the SAN is trying to make a copy after splitting pairs he hammers storage if loud as all the virtual machines on these huge experience LUN disk wait and ultimatly unresponsive.

  Environment: -.

  ESXi 4.1

  HDS AMS2300 SAN

  We raise the issue with VMware and HDS ourselves as it seems that the copying to integrate synchronization pairs at SRM startup, it uses the copy default pace that is 100% hammers so therefore the SAN doing.  We did a test only Lun 4 and it was ok, but our test of full scale of LUN 11 caused a breakdown finally.

  We are planning to bring down our MRS. in the smallest of the Recovery Plans that do not have both LUN at a time that should get out us of this, but wondering what other people are doing?  With all LUNS recovery plan? or separate recovery Plans?

  See you soon

  Abandonment of the copy of the PACE at 1 for all pairs of continuous has solved this problem. Check your copy of the PACE on your pairs and if you have issues drop again to 1.

• SRM API: How to specify if the recovery operation must be planned migration or recovery after disaster

  Is it possible using the SRM API to specify whether we want a planned Migration or a recovery during failover?

  In addition, exactly how do we use the APIs AnswerPrompt and ListPrompts? Examples would be appreciated. Thank you!

  If you want to run the disaster recovery, you must use the recovery option (number 0) failover Mode. The SRM 5.8, we introduce a new Mode of recovery option - migration (planned Migration - number 5). This command is not added to the SRM API documentation, but will be updated soon.

  Example:

  PowerCLI > $RPmode = New-Object VMware.VimAutomation.Srm.Views.SrmRecoveryPlanRecoveryMode

  PowerCLI > $RPmode.Value__ = 5

  PowerCLI > $RPmode.ToString)

  > Migrate

  Orders ListPrompts and AnswerPrompt have nothing to do with the implementation of the plan (planned or disaster recovery Migration).

  -Daniel G.

• SRM with virtual disks not replicated RDM Recovery Plan.

  We try to do a test of SRM with a Virtual Machine that has disks Virutal RDM attached, however these volumes are not replicated with RecoverPoint, they exist already on the side of DR, so basically we're already reproducing machine virtual and it's virtual disks, but these 3 RDM are outside it.  The reason for this is because they have massive volumes that had previously been replicated using another product before virtualize us and there is no need to replicate this info again because the volumes are never going to change.

  
  Then...

  It is done within the SRM, when you look at tab groups protection, we see a warning "not configured VMs: 1" because SRM / RecoverPoint believe that this virtual machine has volumes that are not replicated.

  Is there any way to exclude this or saying SRM to ignore what we will manually add the RDM on the DR side ourselves?  Also on the side of DR, I don't see the VM under hosts and Clusters as I see the replication of other virtual machines, it is because MRS. think he isn't entirely reproduced?  Or is it a manual step where you need to create a virtual machine on the side of the DR and to disable this option in order to see it?  Also, we see what VM listed on the virtual computers tab for this group of special protection, but the other two VMs residing on the sam MON appear and I already tested successfully the two machines.

  Here are the settings in the file for the RDM .vmx

  scsi0:4.filename = "server_3.vmdk".
  scsi0:4.mode = "persistent".
  scsi0:4.ctkEnabled = "FALSE".
  scsi0:4.DeviceType = "scsi-disk" hard
  scsi0:4.present = 'TRUE '.
  scsi0:4.redo = «»

  scsi0:5.filename = "server_4.vmdk".
  scsi0:5.mode = "persistent".
  scsi0:5.ctkEnabled = "FALSE".
  scsi0:5.DeviceType = "scsi-disk" hard
  scsi0:5.present = 'TRUE '.
  scsi0:5.redo = «»
  scsi0:6.filename = "server_5.vmdk".
  scsi0:6.mode = "persistent".
  scsi0:6.ctkEnabled = "FALSE".
  scsi0:6.DeviceType = "scsi-disk" hard
  scsi0:6.present = 'TRUE '.
  scsi0:6.redo = «»

  Hello

  You do not see the DR VM (usually it's called shadow or the space reserved VM), because currently the VM is not protected by SRM. You can't manually create.

  What you can do is run the Wizard "To configure Protection" for your virtual machine. One of the steps, "storage options" or something like this, allows you to select specific disks and mark as detached, allowing you to protect the computer with disks not replicated virtual.

  Michael.

• Recovery test intervals

  I have run tests SRM of stimulus packages on a weekly basis. I also ran a real recovery and failback for additional checks.

  I wonder how many times the second type of test must be performed annually, semi-annual etc.

  What do you and recommend in terms of test frequency?

  A true recovery must be performed for the test when Vcenter/SRM is upgraded or performs sufficient tests?

  It's great that you test so frequently. Nice to hear. It really depends on your environment and your needs. When I was a customer and a partner helping customers I find generally that it is once every 6-12 months, if it was really more to keep the company happy, from my point of view, a test of the stimulus package was enough.

  A failover/failback would make sense after the upgrade, although it would not necessarily be the entire plan. It could just be done for a subset or test group.

  As I said at the beginning, although it really depends on your comfort level and needs of your business.

• QUESTION: Disaster recovery for Test db with RMAN

  Hi friends,
  
  I'm working on ORACLE10g on RedHAt AS4.
  I'm doing on my test database using the catalog of recovery after a disaster, and I deleted all the contrl / redo n /datafile of the system parameter file, when I try to conneciting to Rman I got under question.
  
  
  [rman@linux1 bin] $. / target rman.
  Recovery Manager: Release 10.2.0.1.0 - Production on Mon 28 Feb 11:58:11 2011
  Copyright (c) 1982, 2005, Oracle. All rights reserved.
  connected to target database: RMAN (DBID = 1629436196)
  RMAN >
  RMAN > set dbid = 154326135
  executing command: SET DBID
  RMAN-00571: ===========================================================
  RMAN-00569: = ERROR MESSAGE STACK FOLLOWS =.
  RMAN-00571: ===========================================================
  RMAN-03002: failure of order CSET at 28/02/2011 11:59:13
  RMAN-06188: cannot use command when it is connected to a mounted target database
  RMAN >
  
  
  My problem is... I m unable to set the DBID of my prod db, RMAN (catalog) connects to itself as the target of DB,.
  
  Any guys help?

  Looks like you have connected only to the catalog database.

  do not forget to target = database to restore

  so you should

  (1) set ORACLE_SID = db restore
  (2) rman
  target connection.
  Connect catalog...

• Initial (and probably more common) question of Dr on SRM and private networks

  Hello

  Can anyone share best practices on the ground with regard to my question archiecture of network:

  According to the guide of the SRM administrator: Mrs. in protected sites and recovery must have access to the same private network (logic).

  We will have protected sites and located on the same ground but both recovery are far from each other on separate parts of the campus and there is network connectivity between sites.

  What are the different ways we can get two separate sites to share a private network?  What is the best practice?

  (1) fiber optic cables directly between sites using Layer 2 switches?  Any potential problem with switching over a distance?

  (2) with Layer 2 VPN MPLS? Enough to handle the traffic scenarios and failover/failback mature?

  (3) should work around SRM to allow a routed 3-layer solution?

  I'm from VMware

  As I said more simple solution is scalable layer 2 making available of VLANS / subnets in two sites in order to avoid the re-ip address of the recovered virtual machine. MPLS and direct DC of interconnection on fiber/xWDM were popular alternatives for the extension of the L2. As with anything there are advantages/disadvantages for things like network and mobile IP service (for example, routing, load balancing, etc.) of persistence or rigidity. The Cisco & VMware following white paper might be of interest that it focuses on a new DCI solution that introduces new options for VM environments. The paper is mainly centered around the Cisco Nexus product range and the ability to now support active/active balancing on various, redundant DC links:

  http://www.Cisco.com/en/us/solutions/collateral/ns340/ns517/ns224/ns836/white_paper_c11-557822.PDF

  You will notice the support of max distance of 200KM that is referenced in the document comes from the limits of support for the underlying storage replication (200KM for most synchronous suites), but there is also the effect of the reactivity of latency and VMotion application considerations. Site Selector Global (Cisco GSS) can be used for the traffic of a smart-balancing the load on the front-end server application as well as for unloading DNS service, DHCP and other common services - it would be more of a supplement to the core DC interconnect solutions mentioned above and not necessarily an exclusive alternative.

  Not sure what you mean by a workaround for Layer 3. SRM has the ability to re - ip of the VM during recovery, allowing you to change all THE info from the network stack inside the VM (network/gateway/dns servers ip/mask) and even generate scripts of ddns, but according to my post previous changing network info is in almost all cases, I worked with a customer "2nd choice".

• How to designate VM as not critical in 6.1 SRM

  As in earlier versions of SRM, it was possible to designate VMS in recovery site as non-critical, how this can be done in srm 6.1?

  Hello

  Recovery test plan steps step 3 is suspends criticism VMs on recovery site. Right-click on it and you can call the dialog box add the VM not critical. By recovert not necessary step is number 8. See the screenshot.

  Kind regards

  Nik

• Publish with seizure of writable on the recovery storage

  Hello

  I received the following error message when I try to recover a virtual machine, Windows 2012 R2 on a 100G data store and an attached physical RDM mode:

  8. change of recovery to writable storage Site	-Could not retrieve the datastore "SRM-TestDS. VMFS volume residing on recovered devices "" 52"" cannot be found. Recovered machine '52' not found after rescan HBAS.	2014-09-08 21:19:14 (UTC 0)	2014-09-08 21:19:24 (UTC 0)
  8.1 protection group SRM-Test-FO	-Could not retrieve the datastore "SRM-TestDS. VMFS volume residing on recovered devices "" 52"" cannot be found. Recovered machine '52' not found after rescan HBAS.	2014-09-08 21:19:14 (UTC 0)	2014-09-08 21:19:24 (UTC 0)

  The establishment's SRM 5.5.1.8569, with IBM SVC on sites running SRA version 2.3.0.  The RDM presents itself very well on the recovering site, but the data store the virtual machine is on gives the above error.

  Any help would be most appreciated.

  Thank you

  Thom

  I solved this problem today.  It turns out that the storage should not be mapped to the ESXi hosts on the recovering site.  Guide the user to the SRA had only to a requirement of configuration.  Once I deleted the mapping drives, cleaned the mirrors and plan and groups protection recovery and re run the test that it worked fine.

  Thom

• How to reverse the primary and recovery sites?

  I have a RS v5 working environment has been installed a year ago.  Failovers and test failovers work well.   However, now I have said that I need to reverse the primary sites and recovery, as the roles of our data centers have changed (GRRRRRRRR!)

  I'm under ESXi and vCenter v5.  I have a vCenter server and SRM server in each data center.  I have multiple LUNS 500 GB with protected VM.  How do I go about replacing this SRM such as SiteA (primary) infrastructure / SiteB (recovery) becomes (resumed) SiteA / SiteB (primary)?  Should I rebuild completely?  Or can I keep the existing installation on the servers of the SRM and change the configuration of SRM in vCenter?

  Thank you!

  Why not simply perform a migration scheduled from A to B and then re - protect?

  Rym.

• SRM with vSphere Enterprise and vSphere Advanced

  Hello

  I saw the FAQ license RS saying that SRM requires no vSphere all be of the same edition and features will be affected by which edition is used.

  Which raises a number of questions in my mind.  For example if my environment is as such:

  • Production site with multiple ESX using Enterprise Plus with Nexus 1000V
  • Recovery site with several ESX using Advanced edition
  • Enough to set up for two sites license SRM and vCenter

  I guess that network should not much question, as I can manually map the port network group on the recovery plan.

  So my questions:

  1. Can I configure manually this ESX made uses SRM for commissioning virtual machines running the recovery plan?
     • I don't have a DRS on the DR site to distribute virtual machines and use all the resources.
  2. What happens when SRM tries to start a virtual machine with more than 4vCPU in the DR site?
     • Advance Edition limited my VMs to 4vCPU

  See you soon,.
  Isabelle

  Your correct indeed neglected the advanced part, DRS is not supported. But another question for you: do you only use the equipment have the Dr site for DR purposes or you also make use of the material for other purposes as for example the test? Is it not that you can always create a business case to buy a license of the company and use the site of recovery for test purposes. In case a disaster occurs and a failover is invoked, you can add stimulus to stop your test virtual machines on the recovering site, before the SRM protected VM. This way make you full use of all the material on the protégé and the recovering site.

  Looking at your question again, it makes me think. If you create a cluster on the recovering site and map this cluster cluster on the protected site. All virtual machines that are protected will be started on the cluster during failover. OK you won't have DRS so resources are not distributed on the hosts, but they run and computers vMotion virtual host to another, you can manually.

  Point which is that you want to map individual VMS to ESX host, making your your design SRM very complex and difficult to maintain. Rather go for mapping clusters and take account of the fact that you have to manually virtual computers vMotion between host computers to create a balanced distribution of resources (or buy some DRS). It makes your life a lot easier.