[tech] More unhappy storage

Andrew Adamson bob at ucc.gu.uwa.edu.au
Mon Aug 5 23:22:07 WST 2013


Evening,

Another storage device we have behind the scenes that you probably never 
hear about is a NetApp FAS-2020 NAS. It does storage of /services as well 
as motsugo's /vmstore. We got donated this directly from NetApp a couple 
of years ago and we love them lots. It has dual controllers called onetel 
and nortel, and they can be controlled via their web interface [1] or over 
ssh. I'm telling you this because I looked at the UCC wiki only to 
discover we have zero documentation on them.

Anyway, in the latest electrical assault on UCC that also broke the SAN, 
another disk in the NAS had an amber light flashing. We've had a flashing 
light one disk since day zero that we haven't really worried about, but 
generally amber lights are a bad thing so I investigated further. The 
status on both the webpage and from `sysconfig' on the command line 
indicated that everything was fine - no failed disks and all the 
'aggregates' (collections of disks in RAID 4) were fine. A disk report on 
the web interface didn't show any failed disks. However, a disk count of 
registered disks on the system came up two short.

Further investigation with `sysconfig -d', which gives disk details, 
revealed that nortel wasn't registering that the disks were there, which 
leads me to believe the FAS-2020 doesn't tell you about disk failures 
after a reboot. Just something to watch out for!

All this is not particularly concerning as nortel was set up with several 
hot spares for just this scenario. We still have double parity and a hot 
spare, however from what I can tell we need another spare disk if we want 
the controller failover to work (onetel can take over from nortel if 
nortel dies). We got some 300G SAS disks with the latest gear from Apache 
Energy, and according to [2] we should be able to flash firmware on them 
that will allow us to use them as NetApp disks. Failing that we might be 
able to approach NetApp for some assistance. 

If anybody is keen to learn more about the NetApp, this is a really good 
piece of enterprise kit to know about, and I would be happy to give a 
quick crash course and hand over the disk replacement task to somebody 
with more time. I found [3] to be a particularly helpful resource for 
getting up to speed quickly.

Andrew Adamson
bob at ucc.asn.au

|"If you can't beat them, join them, and then beat them."                |
| ---Peter's Laws                                                        |

[1] from within the ucc network: http://nortel.ucc.asn.au/na_admin
[2] http://www.liveinternet.ru/users/vardomskiy/post127616861/
[3]https://communities.netapp.com/servlet/JiveServlet/previewBody/2999-102-1-3620/NetApp-Basic-Concepts-Quickstart-Guide.pdf



More information about the tech mailing list