#Will these different FW's be included in the ONTAP UPGRADE?
1 messages · Page 1 of 1 (latest)
Disk qualification, unless something has changed RECENTLY, has never been included directly in ONTAP. Why?
I suspect It’s more or less built in with each release, as far as I know. Think of this, when ONTAP is built instead of pulling the disk qualification package, it pulls the data from the file and is compiled into ONTAP so it knows, however, you should always have the latest file installed anyway. Why do I say this? A long time ago (7-mode) disks would show up with a big yellow tag that says to either have a specific release or later installed or install the latest DQP.
Therefore I surmise that it’s built into ONTAP based on the date of build. Either way, always update.
Now, the other files. It takes about two weeks for ONTAP to be built and tested. Typically, the files that are available at the time ONTAP is built (not released) may be included. If there is more of an urgency to get the p release done, the sp/bmc may not be included.
Always grab and install the disk/shelf/sp-bmc firmware. Non disruptive. Just do it
The bios/loader is included in ONTAP as needed. Just go to the download page for SP/BMC/BIOS and look at the support matrix. It tells you which release things are in
to elaborate on this, there were disks back in the day, I think the first 2TB disks or something like that, which would be detected as 1TB disks if you didn't have the correct DQP file installed. If you then (blindly) added your new shiny disks to an existing aggregate, you just wasted half of your capacity with no way back. So yes, at least back then, having the newest DQP was quite important 🙂
Agreed. I tell people to download and install the DQP, Disk FW, Shelf FW and SP/BMC FW at least every three months. If nothing changed, no harm, nothing happens. Keeps those up to date as possible. Especially after ONTAP 9.8 where all those can be installed with the GUI…why not keep up to date and safe?
Could not agree more!
Always grab and install the disk/shelf/sp-bmc firmware. Non disruptive. Just do it
Until the day comes again when it isn't. I have seen a filer pair outage triggered by a drive firmware update. A long time ago, but it did happen.
You can btw check here if a new firmware is included in a certain ONTAP-release: https://activeiq.netapp.com/system-firmware
Right...a LONG time ago. A quick bit a history...I was personally at NetApp when the first "HA" pairs came out. I managed to take down the entire HA pair with a disk firmware upgrade. Evidently, they engineers did not properly fence the process and when one node started the process, the other node did not like it an they actually both panic'd.
TODAY: when left to auto-updating, ONTAP will only do one disk at a time. If the system does not have any spares, I think the process actually terminates. That way if there is a problem, it can at least rebuild.
So we can blame you for making disk fw updates work as well as they do? 😄
https://docs.netapp.com/us-en/ontap/upgrade/concept_when_you_need_to_update_the_disk_qualification_package.htmlou need to download and install the DQP in the following situations. A best practice is to also update the DQP regularly; for example, every quarter or semi-annually.
Whenever you upgrade to a new version of ONTAP.
The DQP is not updated as part of an ONTAP upgrade.
So...yes it needs updated.
FWIW...before Dos Equis came out...(ONTAP 5.2 = Dos Eqious -> Two Beers, Two Filers) I was invited as one of about 6 field SEs to test out HA before it was implemented in the field. I personally found 3 "show stopping" bugs that had to be fixed before shipping to customers. I just wish I had found that Disk FW update issue before going public.
I'm glad that you found the bugs too - we found the fw update bug in the wild (on a 3140 IIRC). But thanks for find some of the other ones... And all y'all can thank for me running 9.4RC1 in production in a very large, very critical environment before it went GA. I still shudder at the thought of making the call to do it.
Hah! I used GX in production only to find that the merge from 7-mode was way off. I hit a hardware bug the 7-mode handled fine (soft assigned shelf id on fc). GX did not. It went through the panic, reboot loop twice. In the third time it jettisoned nvram forcing a WAFLiron which took nearly a month…was using GX for coral volumes which never made it into CDOT!
coral, was that the HPO (striped volumes)? if so, then it actually did make it into the first few cDOT versions, but only for people who already had it. When you had the correct license key installed you could however still create and use those for a few versions
Yeah. It was such a bad error, ended up going back to 7-mode (and my Netapp got me something like 70% off for 4x new controllers) which took like 2 months and waited for 8.1. I was kept in the loop about coral volumes and we did hear that it was being removed. So never tried again. Of course Netapp tried infinite volumes which have already gone the way off the dodo. If I were in the same spot again, I'd definitely be used FlexGroups.