Hi all
I've just set up my test SRM biosphere and I have just one little problem! I cann't see any replicated datastores in configure array managers wizard
System set up per site is as follows:
HW
- EMC Symmetrix DMX3 with lun offset set for lun visibility for ESX hosts. Microcode 5771
- SRDF/S with 3 lun in state synchronized.
- 2xHP BL685c G1 with Emulex FC adapters connected to Brocade SAN switch.
SW
rdmssrm98 :
- Win2003 EE R2 SP2; Virtual Infrastructure 2.5.0 Update 3; SRM 1.0.0.-97878; EMC SRDF SRM Adapter v.1.0; EMC Solutions Enabler 6.5.2
bq. - all EMC software is latest available via powerlink.
- SYMAPI Server daemon installed and started:
C:\Documents and Settings\dejan>stordaemon list
Available Daemons ('[*]': Currently Running):
storapid EMC Solutions Enabler Base Daemon storgnsd EMC Solutions Enabler GNS Daemon storrdfd EMC Solutions Enabler RDF Daemon storevntd EMC Solutions Enabler Event Daemon storsrmd EMC Solutions Enabler SRM Daemon storstpd EMC Solutions Enabler STP Daemon storsrvd EMC Solutions Enabler SYMAPI Server Daemon
- netcnfg line: rdmssrm98 - TCPIP rdmssrm98.sit.pbz.hr 10.100.101.19 2707 SECURE
rdssesx14:
- ESX 3.5.0-110181; EMC Solutions Enabler 6.5.2
- created 3 datastores each on its own lun. luns are visible on both esx servers. of course, on this server are type RDF1 RW and on the recovery side are type RDF2 WD
- 6 VMs
- firewall configured with a "hole" for SYMAPI Server on port 2707
- edited global path variable for /usr/symcli/bin
- SYMAPI Server daemon installed and started:
root@rdssesx14 root# esxcfg-firewall -query | grep SYMAPI
SYMAPI : port 2707 tcp.inroot@rdssesx14 root# stordaemon list
Available Daemons ('[*]': Currently Running):
storapid EMC Solutions Enabler Base Daemon storgnsd EMC Solutions Enabler GNS Daemon storrdfd EMC Solutions Enabler RDF Daemon storevntd EMC Solutions Enabler Event Daemon storwatchd EMC Solutions Enabler Watchdog Daemon storsrmd EMC Solutions Enabler SRM Daemon storstpd EMC Solutions Enabler STP Daemon storsrvd EMC Solutions Enabler SYMAPI Server Daemon
HINT1:
When I edit netcnfg to point to symapi server on esx host, I get an error executing sym commands (bla bla) as sson as I click on rescan arrays button in Configure Array Managers.
But, when netcnfg points as stated in system setup above, rescan works for some time but it doesn't show any datastores. The list is empty.
Log file is attached.
HINT2:
Why SRM cann't execute rescan arrays when I specifie ESX host as SYMAPI server? maybe there is some foe around here because I use my SRM box for DMX operations.
Please help because I'm going crazy around here. Tnx in advance!!
You need to have the SPC2 flag enabled on the Symmetrix front-end adapters serving your ESX hosts (both protection and recovery site).
Please see this post:
Also, on EMC Powerlink please look for the white paper entitled:
"Enabling SPC-2 Compliancy on EMC Symmetrix DMX Devices Connected to VMware VI3 Environments"
for details on what this flag means and the implications of enabling it. You can verify that the flag is not set by running the command:
symcfg -FA ALL list -v | grep SPC
to list the value of the flag on all FA's (or replace ALL with the FA number that your ESX hosts are connected to, if you know it)
Essentially, SRM is attempting to match the WWNs of the devices discovered by the SRDF adapter with the WWNs of the LUNs that the ESX hosts see through a SCSI inquiry. The EMC adapter returns LUN WWNs in NAA version 6 format; the SCSI inquiry will return NAA 6 format only if the SPC2 bit is set. Otherwise, SRM is not able to match the output of the adapter with the storage ESX hosts see, and will return "no replicated datastores".
-jasonb
Yeah, I saw that thread before posting my post but I didn't had info on that. I am still waiting confirmation from my storage guys.
My vmware contact, with whome I'm trying to build this lab at my environment, sent me this new data:
- devices must be in a device group
- device group must not contain devices not being used by SRM
- device group must contain bcv or timefinder device for SRM to use during testing
ok for DGs, it makes sense because when executing symrdf failover you need to specify device group.
But, I don't, and I cann't have, bcv or timefinder device in device group. Only rdf1,2 compliant devices.
back again
SPC2 is enabled per wwn (symmask set hba_flags) and still no luck.
this is output from "symmaskdb list database -v"
Originator Port wwn : 10000000c96f9f92
Type : Fibre
User-generated Name : rdmssrm98/p1
Visibility : No
FCID Lockdown : No
Lun Offset : Yes
Offset Value : 00ec
Base Value : 0003
Heterogeneous Host : No
Port Flag Overrides : Yes
Enabled : SPC2_Protocol_Version(SPC2)
Disabled : N/A
Dynamic Addressing : No
Authentication State : N/A
Devices : 06BB 06C3 06CB
Hi,
>> SPC2 is enabled per wwn (symmask set hba_flags) and still no luck.
You may need to rescan ESX hosts on the protected site to pick up this change on the array. If this doesn't help post SRM logs from the protected site so we can take a look what's going wrong.
-Masha
thanx for reply divintas
I read it about 15 times and even in reverse
it is so frustrating that you cann't find all the requirements in one place. they are all scattered around in white papers, knowledge bases, user experience...
mariab, I didn't do rescan hba, I rebooted the damn thing :smileycool:
and of course still no luck
SRM logs are attached.
Hi dpetrovic,
From the logs it still appears that ESX doesn't see WWNs of the devices correctly. Usually this is due to SPC-2 or SC3 flag not being enabled. From your earlier posts it sounds like you've checked SPC-2 flag and it is enabled. However, I compared the log from your original post and the log from the latest post and I see no difference in how ESX sees the devices. If possible, could you rescan ESX host, then re-create datastores on replicated devices and try configuring SRM again. If this doesn't help then please attach new logs and I'll try to look again.
-Masha
done and I got a new set of logs for you
can you please explain a litle more that part that ESX doesn0t see WWNs of the devices correctly? because, there are statements in log file that he discovered LUNs and replicated devices.
you should also know that lun offset is implemented on this LUNs.
SPC-2 and SCSI3 flags are enabled from the start. SPC-2 is enabled per wwn and not per whole port.
Many tnx in advance!
Hi dpetrovic,
>> can you please explain a litle more that part that ESX doesn0t see WWNs of the devices correctly?
ESX sees 4 devices:
device WWN
-
vmhba0:0:0 31:30:30:33:38:33:30:38:32:30:30:30:53:59:4D:4D
vmhba0:0:6 31:30:30:33:38:33:36:42:42:30:30:30:53:59:4D:4D
vmhba0:0:7 31:30:30:33:38:33:36:43:33:30:30:30:53:59:4D:4D
vmhba0:0:8 31:30:30:33:38:33:36:43:42:30:30:30:53:59:4D:4D
These WWNs are not correct because WWNs should start with 60:06:...
>> SPC-2 is enabled per wwn and not per whole port.
"Enabling SPC-2 Compliancy on EMC Symmetrix DMX Devices Connected to VMware VI3 Environments" white paper available through powerlink says that SPC-2 support can be turned on for each FC port or on a per-initiator basis. Which option did you choose?
-Masha
ups, we didn't enabled SPC-2 on initiators for ESX hosts.
we needed 3rd pair of eyes to figure out that one because we were looking at it and didn't saw it
we can now see datastores and succesfully created one protection group.
thnx for your time and effort!
for everybody out there - RTFM!
for everybody out there - RTFM!
Hey, thanks for the free advert!!! 😛
Regards
Mike