11
Jun
09

Upgrading – But wait… there’s more!

Today’s Issue:

Restoring a UCS archive from 9.3.1 configuration to a 10.0.1 upgrade installation on an F5 BIGIP LTM 6400

Where We Are:

Through various iterations, have “successfully” upgraded to 10.0.1, but without the transfer of settings & configuration objects (virtual servers, nodes, pools, user accounts, etc)

How We Got Here:

Continuing from yesterday’s post:

Test 1 – Change the hostname to match the previous… worked like a charm.

Test 2 – import & install of the old version to the secondary partition was actually SIMPLE with the new interface.  So far has taken about an hour for the new install to “activate” reboot…  once that comes up, make sure the UCS config is still there, or re-apply the archive (thinking re-apply, as this will be on a new partition without the old files/filesystem).

Don’t know what happened here, but hard reset after waiting about an hour.  May not have been a great idea, but it did come back up in 9.3.1 mode, was able to restore the UCS (which even brought back the user accounts), and all looks OK.  Rebooted again to make sure it came up as the default install version, and all seems to be well.

FYI… nice little note:  if you blow the password that you changed on the admin account, and can only get in as root (i.e. you can only SSH in, apparently root can’t access the GUI), use this command to change the admin (or potentially any other account) password:

F5passwd <account_name>

Another lingering issue experienced yesterday, was the notion of redundant power supplys, and if they actually are… there are 2 plugs in the back, and it won’t spin up without both of them plugged in.  As well, if you happen to kick one of them out while it’s on (like I did, mid-upgrade – scary moment), the thing takes a nosedive.  How is that redundant/failover power?  Do I actually have toi configure that somewhere?  We’ll look at that more, later.

UPDATE: it seems that either one of the power supplys is dead, or not installed, as you don’t get anything if it’s the only one plugged in, and if you remove the other, system dies an instant death.

Ugh.

The upgrade process that we seem to have ironed out, now doesn’t show any interest whatsoever in running.  Following the guide we created while in process last go-round, we seem to die a quick death on the first command:

im BIGIP-10.0.1.283.0.iso

Looks like it can’t write to /tmp for some reason (logged in as root for chrissake!!)

So all I want to do is get rid of the 10.0.1 install… the switchboot utility will allow you to select which version you’re booting to, but won’t let you kill the un-wanted one.  For that, sounds very simple, run this little script called “local-install” and it walks you through it.  Only problem is, it doesn’t appear to be there, so I now need to download it (438MB) and upload to the device, then run it.

Hmmm…

So we ran the local-install to remove the 10.0.1 version from the 1.1 HD install slot, rebooted, and confirmed it wasn’t there.  Re-pushed the BIGIP-10.0.1.238.0.iso to the device into /shared/images, ran the IM… command, then the image2disk… command (left out the –nosaveconfig switch) which ran REALLY fast (15 seconds as compared to a few minutes), rebooted once, and came right back up not only with 10.0.1 installed, but all configs and user accounts intact.

I’m a little nervous…

Advertisements

1 Response to “Upgrading – But wait… there’s more!”



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: