This post is more than 5 years old
4 Posts
0
2093
vSphere 5 SRM 5 SRA 2.0 for Recovery Point
Hi everyone,
Hope this is the right place to post.
Basically we have 2 sites 2 san's 2 RecoveryPoint appliances.
Problem:
When we initiate a SRM test for a particular recovery group the test fails because it can not see the datastore. I am pretty sure the SRA is passing the commands to recovery point but the datastore never shows up. So the test fails...I wonder if anyone can help me walk through the stuff manually so I can find out why its stopping...
emc support told us that we should change the SRM datastore rescan to greater than 20 minutes I just dont buy it.
can anyone point me on what to check for in RP or how to track this one down. It's seems like its not creating the snapshot,journal,bookmark (not sure what its called here)
Main Site:-
- VNX5700 has flare code 5.31.000.5.006
- RecoverPoint 3.4.2
Dr Site:-
- CX4-120 4.30.000.5.522
- RecoverPoint 3.4.2
vSphere 5 SRM 5 SRA for Recovery Point 2.0
HELP!!!
-Eid
emashaal
4 Posts
0
October 31st, 2011 04:00
Ok as it turns out there is an issue with ESX 5 and the EMC flarecode. VMware was aware and turned it over to emc integration team.
If you are unable to see your datastore you will see a locking issue in the vmkernal log. You have to turn off hardware assisted locking..(advanced setting, ..vmfs3)
VMware's support (out of ireland) was amazing once finally engaged ...... but since this was a priority 2 case and with time zone issues it took way long time to get some help..
EMC's support was too busy giving us level one techs who want us to reboot stuff and do other mundain motions while their level 3 engineers know about the issue and its solution...really annoying and confirms that if you know your issue ask for esclation from the first call!!!
;-)
emashaal
4 Posts
0
October 28th, 2011 17:00
Do you mean the replicated luns? They are zoned to the ESX host on the DR side.....
I imagine its supposed to happen like this
SRM "hey RP i got a test here create me a snapshot then somehow mount it to that ESX host"
but how did it know which host to mount it to?? where is that info
then the rest is easy
but in my case SRM tell RP but the datastore never shows up in the rescan....
etaljic81
1K Posts
1
October 28th, 2011 17:00
Did you add the target LUN to the ESX Storage Group at the DR site?
etaljic81
1K Posts
1
October 28th, 2011 17:00
You have to add the target LUN to the ESX Storage Group on the target array. That's why it's failing. Create a Storage Group on the DR array, add the ESX host(s) to the Storage Group and also assign the target LUNs to the same Storage Group.
emashaal
4 Posts
1
October 29th, 2011 03:00
Ernes,
Actuall the luns are in the SG for the DR host the problem is that esx 5 does not see any datastores on the lun....I dont know whether its an SRM issue or a recoverypoint misconfiguration issue....
BTW: RecoveryPoint does properly create a snapshot and the esx host sees it.....however i am confused because it seems that sometimes esx 5 see's those replicated luns as detached and sometime it sees them attached....seems like there is no method to the madness....
I can go into esx and add storage and see my lun but if i write a signiture to it it is just an empty lun.....should esx just see the datastore and mount it.....
so much for emc & vmware integration!!
etaljic81
1K Posts
0
October 29th, 2011 18:00
I've intergrated RecoverPoint with SRM many times and it always worked. There must be a setting that's not correctly set in the RecoverPoint Consistency Group. Check to make sure you select ESX under hosts for the CG and also make sure that the CG is set to be managed by SRM.
etaljic81
1K Posts
0
October 31st, 2011 06:00
Very good to know. Thank you emashaal. I'm definitely going to keep this in mind if I come across ESX5 with SRM. Thank you for the points as well