2012-03-13 02:18 PM
We've recently received a brand spanking new FAS2240-4 with dual controllers and fully loaded with 2TB SATA disks. This is a stand-alone filer with no shelves attached.
I'm a little unclear about cabling. Both the NetApp Univ. video I found on setup, and the cabling guide that came with the unit say to cross-connect SAS ports between the two controllers, and to cross-connect the ACP ports across controllers. However, no SAS or ACP ethernet cables were included. The installation engineer said these were not required. Can anyone clear this up for me? I have an amber module fault LED, which may be due to one/both of the above (?).
2012-03-14 12:06 AM
They are not required indeed – single path for internal disks is fully supported. But you are right in that MP-HA is recommended; I can’t comment on NetApp order units structure; I would suggest you raise this issue with your NetApp sales representative.
I do not think fault LED is due to missing cross connection. Usual reason for newly delivered systems is the fact, that cf is disabled. Otherwise just open support case to verify configuration.
2013-04-10 05:11 AM
you might want to check the document provided on the NOW.
There are some cabling diagrams that are there to help you out.
External shelfs on a 2240 seem to be MP cabled...
We just found out that it is single cabled for a project we have in Mexico.
Checking with Netapp why it is not MP.
Hope it helps.
2013-04-10 07:02 AM
The installation engineer may just be right!
2240 systems in an HA pair with no external SAS storage may use single-path HA, which requires no external cabling. However,the recommended configuration is to enable Multipath HA for the internal disks by connecting the SAS ports of one controller module to the SAS ports of the partner.
As per the cabling guide for FAS2240-4:
For fault LED you can try these steps:
1. At the Data ONTAP CLI prompt, issue the “halt –s” command. Please note that this will cause an HA takeover in an HA configuration.
2. Wait for the controller to shut down .
3. On the console port (not RSH, SSH, etc.), issue a “^G”(ctrl G)to switch the console to the SP
4. Log in to the SP.
5. At the SP prompt, issue the “system power on” command.
6. Issue a “^D”(ctrl D)to get back to the Data ONTAP console.
7. Once the BIOS / Loader have booted, issue the “boot_ontap” command.
8. Issue Giveback after Takeover.
For HA configurations, the partner will also need to go through the 7-step process.
2013-04-10 07:51 PM
wow, i never knew that
Learn something every day
The fault light is most likely because there are licensed features which aren't configured or active. i..e if you have CIFS licensed, but don't actually configure CIFS (or start it) the fault light will be on.
2013-04-10 11:19 PM
The SE was there yesterday to fix MP ... yes on a external shelf.
Still I'm confused why he didn't configured it MP from the start? If the functionality is there... we should use it.
The failover (halt -s) look like a windows reboot to fix issues.
I would check "cf status" as aborzenkov described. Otherwise.... "rdfile /log/messages"
2013-04-11 02:44 AM
I agree. And, you need MP for more than one reason. This is the extract from this blog.
Multipath HA for Internal Disks
The FAS2240 is capable of providing MP-HA for all attached disks, both internal and external. MP-HA is the capability to provide redundant paths between storage controllers and disks for dual-controller (HA) configurations and reduces the chance of a failover occurring as the result of path failures. Having a second path to reach storage protects against a variety of possible failures, such as:
•Controller-to-shelf cable failure (in configuration with multiple disk shelves)
•Shelf module failure
•Dual intershelf cable failure
•Secondary path failure (in HA configurations)
While MP-HA increases overall storage resiliency, it also enables Data ONTAP to perform maintenance activities such as disk firmware upgrades nondisruptively. Performance can also improve because MP-HA provides twice the bandwidth to your storage.
1. Yes, please check 'cf status' as aborzenkov described.
2. Also check out this, as indicated by shane bradely.
Error Light on Chassis, but no errors in /etc/messages
3. Another thread, discussing system fault led.