Hi,
i have a question one of my customers is trying to see if they can have SRM (vSphere 5.5) configure several IP addresses in the IP settings in SRM. at present SRM allows me to pick up the current IP settings for the protected VMs and then allow me to type in the new IP for the recovered VM.
The VM in question has several IPs using one NIC, any idea how this is overcome in SRM?
thanks for all help in advance.
Kabir
Hi,
This might be a silly question, but after my initial pick of place holder datastores, I want to change them, but when I click to select them the original datatstores I selected are greyed out....so I cant deselect them?
I can add more to be placeholders but I cant remove the ones currently selected? Do I have to disable replication first or something?
Cheers,
Bilal
Hi Guys,
Basically normally when you use vsphere replication, if you lose the VRMS then you cant recover your vms.
"With the VRMS appliance offline, it is not possible to recover VMs with vSphere Replication. There are some third party articles out there that suggest it is possible to manually alter the files at the target location that make up the replica and recover a VM. While this is possible, I do not recommend relying on that recovery method as it is a very manual process and, more importantly, not supported by VMware "
vSphere Replication Appliance Failure Prevention and Recovery | VMware vSphere Blog - VMware Blogs
Am I right in thinking that when using SRM with vsphere replication, this isn't an issue as you can 'force' the recovery with SRM ........or is this still going to be an issue?
This stems from a question iI posted in Re: running 1 primary replication appliance and running 1 add-on appliance
Cheers for the help guys!
Hi Guys,
I have setup snap mirroring, and I have added the correct SRAs and it is showing as replicating...BUT when I try to move a VM to the replicating datastore...or even create a folder on it at the source site. It fails with Call "FileManager.MakeDirectory" for object "FileManager" on vCenter Server "vCenter2008.gotham.com" failed.
This screams to me a rights issue. I have configured the rights using the web-client (OnCommand System Manager), if I give the hosts Root access to the LUN then SRM/SRAs wont see the LUN and use it for array replication, its in a couple of articles. The hosts need read/write access only for it to work.
But even with the hosts having read/write, I cant create anything on this LUN at my source site.
The LUN at the Test/DR site is mounted as read-only on those hosts as recommended in other posts and is set to restricted on the NetApp side
As it stands...it appears ABR is working (according tot he webclient, it is showing outgoing array replication), but it isn't actually replicating anything but empty space?
snapmirror status
Snapmirror is on.
Source Destination State Lag Status
NetApp1:vol1 NetApp2:vol2 Snapmirrored 00:00:00 In-sync
I am still new to Netapp/snapmirror/NFS in general. This may sound really stupid, but the replicated datastore, has the VMS on it that you want replicated and they are replicated byte for byte as everything on the array is sent across tot he DR site?
Thanks for your help guys!
Bilal
Hello,
I have inherited a vSphere 5.1 environment with SRM 5.1 + Some Dell Compellent sans (2 in the UK and 2 in France).
My question is that we have a number of RDM's included in the replication schedule which I would like to stop SRM from seeing as it produces an error from within the SRM plugin (device "RDM" cannot be matched to a remote peer device).
Datastores that I am replicating and which have VM's on do not show up with the error on the array pairs tab.
Is there a way to tell SRM to ignore these RDM's that I'm not including in any SRM protection group?
In my previous role which was a NetApp environment.. I was able to specify which data store names to look for within the SRA configuration..
Hope this makes sense... if not feel free to reply and I'll elaborate further
Hi All,
our environment for SRM is a non typically one. We have a cold ´recovery site where we recover 3 datacenters. We cannot set an automatically network mapping in the site configuration, because per datacenter we have 30 different VLANs and the VMs are connected to this. It is possible to configure the protection for a VM using powerCLI and set the VLAN ?
We have SRM 5.8 running.
Under certain circumstances, the icon for Placeholder VMs displays incorrectly in the vSphere client.
Let's say you have protected and recovery vCenters, SRMs and vSphere Clients with SRM plugins and DNS all good to go.
When all is good
When you are Windows logged in to your recovery site vCenter (the site containing placeholder VMs) and you launch the vS client, it does not matter how you refer to the recovery vCenter server when launching the vS client, by short name, FQDN or IP, when you then navigate to a container holding your placeholder VMs, the icon for them is "correct", i.e a typical 3-boxes vm icon with the lightning bolt. See first attachment right.jpg.
Now, if you are Windows logged in to your protected (Note: protected) site vCenter and you then launch the vSphere client to login to the recovery vCenter (Note: that's recovery vCenter not protected vCenter), how you refer to the recovery vCenter system affects the correctness of the icon displayed for Placeholder VMs. If you use short name or IP, placeholder VMs display correctly with the lightning bolt icon.
When it's bad
If however you specify the recovery vCenter by FQDN when launching the vS client, remember - you are doing this when Windows logged in to your protected vCenter, then the icon displayed for Placeholder VMs is wrong. See second attachment wrong.jpg. The lightning bolt is not there, but it is also not a standard vm icon. I don't know what the icon is meant to represent.
Go ahead, try it, as I have done on multiple systems. Please read the above carefully when attempting to replicate this.
Is this a bug? Can anybody explain why you should get different icons for placeholders determined by name referral method to the recovery vCenter by short or long name?
Does anybody care? Probably not, but I reckon it's the most important issue facing mankind this microsecond, so your thoughts are most welcome.
Hello
I have downloaded SRM on my computer but I cannot run it directly on the OS.
So I'd like to know if SRM 5.8 have to be installed into an another products (example: Hypervisor + VCenter) ?
Or if I can launch SRM 5.8 can be directly launched on my computer's OS (Windows 7).
Sorry for this (studpid) question and many thanks for your reply.
Dany
Good Morning,
I'm trying to figure out why all of our VMs show up in one datastore group.
We're running VCServer 5.5 with SRM 5.5, and we have identical Nimble Arrays at both sites with array firmware 1.4.11 and the Nimble SRA 2.1.0.
We have about 70 VMs. Each VM we have is in it's own datastore on the Nimble - however we have one large volume that is used for VMWare snapshots and another volume that we use for Windows Paging files. We have three volume collections / consistency groups on the Nimble side - two that take simple storage level snaps and a third that takes VMWare snaps before snapping the storage for replication. I would expect all three of these volume collections to show up as separate datastore groups, however SRM is lumping them together. This morning I created another volume and used Storage vMotion to move the VMDKs that have Windows Paging Files for one of our consistency groups to it's own separate datastore for just those few VMs. Afterward, I asked SRM to recompute datastore groups and they still show up as one large datastore group.
Will the location of the VMWare swap files also combine consistency groups into one large datastore? We try and move the swap files off to other, non-replicated datastores to keep replication traffic down to a minimum.
At this point, having one large protection group isn't causing a major issue, but it might down the road as we haved deemed that some VMs don't need to be protected by SRM, however those VMs are eating up SRM licenses by being part of those datastore groups.
The database for the SRM (5.1) server in our recovery site needs to be moved as it is currently on a Windows 2003 server.
To reconfigure SRM I'm assuming the following needs to be done.
1. Shut down SRM services on both servers.
2. Backup the SRM database on the recovery site database server.
3. Restore the database on the new database server
4. Run a modify Installation of SRM server to configure it to use the new database server.
5. Test SRM
Does that cover it? Am I missing anything (other than taking a snapshot of the server before reconfiguring it)
HI,
I have confgured SRM 5.8 with two sites (PROD and DR) sucessfully. One minor issue i have is that when the Sites are displayed in the web client always my DR site is in top of the list and prod site appears below in the list. And also when i create a protection group by default in the first tab "Name and location" the location is pointed DR - Prod and in the next step "Protected group type" my DR site is by default picked up as protected site and i have to change the radio button to my PROD site and then create it. SRM automatically picks the site order and is there a way to reverse this default order?
Hi there
I've run into a problem and can't figure a solution at all. I'm running SRM 5.8 trial edition for Uni Project work along with HP StoreVirtual VSA creating a Virtual SAN. I'm using SRA 10.5.0.92 which seems to be the correct version required for SRM 5.8 and while SRM is detecting my Replications in vCenter Web Management, it is not recognising a datastore, thus when I come to create a protection group I am presented with the error message:
I definitely have an ESX datastore which is being replicated and the datastore has virtual machines on it.
I've been literally banging my head against a wall for the past 2 weeks-ish trying to figure out what's going wrong.
The SRM logs seemingly indicate that all my storage devices are being skipped under detection as they are "neither iSCSI or FC"
| SRM 5.8 Logs |
|---|
2015-02-17T23:28:07.255Z [02732 verbose 'DatastoreGroupManager' opID=4b543b6f] Skipping extent 't10.ATA_____VB0250EAVER_____________________________Z2ADTEJB____________:3' of VMFS volume '4f050297-02b9807a-2cbb-2c768aadf716:datastore' which is neither FC nor iSCSI 2015-02-17T23:28:07.255Z [02732 verbose 'DatastoreGroupManager' opID=4b543b6f] Skipping extent 't10.ATA_____VB0250EAVER_____________________________Z3T0L3NN____________:3' of VMFS volume '50ad5dea-704efe81-169d-a0b3ccdf2aaf:datastore1' which is neither FC nor iSCSI 2015-02-17T23:28:07.255Z [02732 verbose 'DatastoreGroupManager' opID=4b543b6f] Skipping extent 't10.ATA_____WDC_WD1002FAEX2D00Z3A0________________________WD2DWCATRC128181:1' of VMFS volume '5473e723-b6dc8dbb-4e55-e83935f08fef:PR SAN' which is neither FC nor iSCSI 2015-02-17T23:28:07.255Z [02732 verbose 'DatastoreGroupManager' opID=4b543b6f] Skipping extent 'naa.6000eb354b962645000000000000001d:1' of VMFS volume '54d40a68-8b085f00-ec16-001f2958247c:Server Volume (iSCSI)' which is neither FC nor iSCSI |
The first 3 are correct as they are standard SATA disks but the final entry is my iSCSI target from the HP LeftHand.
Ignoring the last entry which was a test target and currently inactive, the Server Volume (iSCSI) is volume being presented from HP LeftHand, and all the vHosts in the primary group have access to it. vCenter confirms this as it can read the datastore.
I've no idea what else to try and was wondering if anyone has experienced this problem.
Hi SRM gurus,
I have the following situation:
- Trying to remediate a bunch of VMs, which will get new IPs on the recovery site.
- In order to start them all (when disaster strikes), I need to know all the new IPs and update certain text file on one of the VMs with that information (it is linux VM)
what is the best way to do this?
Hi ,
I'm trying to set up a 5.5 SRM environment and I'm having a few problems:
The production site is set up with VC and SRM installed and was working.
When I log on now though I can't connect to VC. ( this is a timeout error and I have reset the timeout in the registry so this may be solved)
On the production site I have installed both the VC and SRM.
But on trying to start VC again I get the error attached and the event log shows the error.
One other issue is that after installing SRM on the reocvery site I cannot download and install SRM from the plug in manager.
The error states that 'it cannot connect to the vsphere client'
Any ideas on this would be appreciated.
Patsfan
I just got SRM loaded and configured. I've tested a few VMs and every test has been successful. However, I have one VM that failed this morning when I tested it.
About my environment:
vSphere 5.5, vCenter 5.5, SRM 5.8
NetApp FAS2040 using SnapMirror
Single VM in recovery plan, approximately 1 TB in size
The problem:
I recently moved a file server to a LUN on the NetApp and began utilizing SnapMiror. I let the SnapMirror replicate over the weekend and first thing this morning I built the recovery plan and executed a test. The test failed after 1020 seconds on the Power On step. Looking at the tasks for the VM in question, it said "Power On virtual machine" and hung there ate 44% before failing after 20 minutes.
Any idea on where to start troubleshooting? The tasks and events for the SRM host only reports "An error was recevied from the ESX host while powering on VM SERVER_NAME". I checked the storage logs as well to just rule out any storage trouble, but there were no errors there.
Thank you
Hello. We have vsphere 5.5. We are looking at doing a poc with srm 5.8 or 6 which is supposed to be GA in the next month or two.
I'm looking for ideas on how to handle updating DNS when we fail a protection group over to the recovery site and re-ip the recovery vm's. We have a replicated MS ActiveDirectory that is in the recovery site.
It's nice that srm 5.8 has the IP Mapping built in now so we don't have to map IPs via csv but not sure of the best way to handle the dns updates. We are going to be doing some testing but I'd like to hear what others are doing as well.
Thanks for any info you can provide!
Jamey
Hi ,
I want to test a recovery plan and I have everything (hopefully) set up correctly.
Both sites are connected and I have configured the Mappngs , Folders , Placehoders and Networks etc.
I've even done a successful fail over with a Protection Group that had no VMs in it.
And therein lies the problem.
When I go to create a Protection Group (using vSR) I don't see any VMs listed.
My VMs are sitting on a Lun on the Production Side and this Lun is mapped to a Lun on the Recovery Site.
Both Luns are Openfilers
Can anyone tell me why I can't see the VMs or for that matter the Lun?
Is this because the Lun is Openfiler?
This is my first go off of SRM so don't have much experience with this.
See attached PS.
Many Thanks
Patsfan
Testing my new SRM 5.8 installation, and by testing I mean that I am actually running a "Planned Migration" on a recovery plan using a test VM that was operating in my test environment. The failover from my protected site to my DR/remote site worked flawlessly. On the reprotect the wheels fell off:
First error:
Protection Group '<Protection_Group_Name>' has protected VMs with placeholders which need to be repaired.
As the vmware kb states I re-ran the reprotect again with the "force cleanup" option and the plan progressed further. Until...
Second error:
Placeholder VM for the protected VM 'VM NAME' is missing. Please restore placeholder before continuing.
I could not get past this error and could find little to no documentation of this error.
Now a little bit more about my setup. I am using Vsphere replication and only have replication going from my protect site out to my remote/DR site.
Since I could not get anywhere with the issue I called support and the VMWare engineer told me that in order for the reprotect to work that I had to setup a new Vsphere replication going from my newly failed over, vm at the DR site back to my source/production site?
HUH? Is that really true and if so then what exactly is SRM doing for me? I was under the impression that SRM would automate most of this for me but if I have to setup a special one-time custom replication just to failback then I really don't understand the point of even having this supposed reprotect option. It would be just as easy for me to setup a new one time recovery plan going the other direction.
Can someone clarify this for me?
We recently upgraded to SRM 5.8.0.1.
When performing a test recovery of our recovery plan, I see a lot of VMs that fail to recover with "Error timed out waiting for VMWare tools" or "Timed Out waiting for Customization." I realize I will probably need to bump the timeout values for VMWare tools startup, however what is really weird is that since the SRM 5.8 upgrade whenever I do a test recovery - I see most of the VMs in all priority groups (1-5) time out at once - as if the 300 second timer started for all priority groups as soon as priority group 1 started to power on VMs. Has anyone else seen this? I"ve ran two test recoveries, including one where I rebuilt the protection groups and recovery plans from scratch.
Hi
I'm trying to build a project with features of SRM and I would need to know if you can get full features of SRM inside a vcenter essentials +.
At location A you have 2 hosts (a & b) and 1 san.
At location B you have 1 host (c) and another san.
I want to live replicate a and b in c while the relevant data inside the first san will be replicated to the other san, in case of a disaster in location A.
Well, it's the first mean of SRM but as vCenter essentials + does not contain vstorage, do you think the replication feature will function with online virtual machines ?
I didn't get a clear answer from pre sales support, and technical support sent me to this place...
Hope I'll get some answer.