[SOLVED - manual upgrd] Upgraded to 2.1.4 and now my WG Firebox X700 won't boot!

phatty

My X750E unit suffered from this same bug…. Luckily I have 2 units in high availability mode so I didn't have to come running into the office last night when the unit didn't come back to life on its own.

So I was able to get it to boot off the other 2.1.3 image and then cloned my 2.1.3 image to the other boot partition until I see confirmation that the update works properly. I was running the 4g version FYI.

m4f1050

I'm with you on the wait…. If anybody is nice enough to let us know when this issue is fixed we will be even more grateful than giving us a super awesome firewall solution! Thanks pfSense!

stephenw10

I very nearly vpn'd into my home box to upgrade it but I think I'll hold for a bit. The box is in London but I'm currently in Texas and I don't fancy trying to talk anyone through a remote slice switch.
Keep an eye on the mailing lists. Nothing there yet and if this really is going to nano_vga I would expect it to bite plenty of users.

Steve

bennyc

So I'm screwed. Exit the FW from the serverroom, it's on the desk now >:(
Checked fast the CF, I still see the 2 partitions. (so my 2 slices are still there)
Booting the box, I only see output in 115200 baud. Nothing shows when I set it to 9600.
But if it upgraded to the vga image, it kind of explains why I get no serial output I suppose?

@phatty, how did you get it to boot the alternate slice without serial output??

As it is already on the desk, the easy route would of course be just flash the latest image… but there's little fun in that, so I'll continue the search for now ;D

@steve; it seems that alix is also affected (if I understood it correctly), see post from "xbipin" earlier in this thread.

xbipin

yes alix affected only regarding the console, the rest still updates and reboots fine also the rest works also

bennyc

Okay, tnx for the feedback, and good that it comes up.
That might explain why others don't report issues (I like to assume I'm like most, and I don't look at the console unless there's issues ;))
Making no progress here with that darn CF. Any hints on how to safely mount that BSD partition in windows?
(sorry, bsd noob here :-[ Still learning…)

phatty

@bennyc:

So I'm screwed. Exit the FW from the serverroom, it's on the desk now >:(
Checked fast the CF, I still see the 2 partitions. (so my 2 slices are still there)
Booting the box, I only see output in 115200 baud. Nothing shows when I set it to 9600.
But if it upgraded to the vga image, it kind of explains why I get no serial output I suppose?

@phatty, how did you get it to boot the alternate slice without serial output??

As it is already on the desk, the easy route would of course be just flash the latest image… but there's little fun in that, so I'll continue the search for now ;D

@steve; it seems that alix is also affected (if I understood it correctly), see post from "xbipin" earlier in this thread.

Benny, the bios output, and the pfsense output for my setup is 115200, but at 9600, for a very short second, I was able to wait and eventually see the boot options for which slice to boot off of.. It took a couple tries for me to time my button pushing just right to get it to boot off the older image.

Supermule

This is exactly the reason why I run it in VM's…

KOM

This is exactly the reason why I run it in VM's…

+1.

I can't be bothered playing around with finicky hardware, CF cards and all that nonsense.

Supermule

No if uptime is important, then I see no alternative….and its a lot more flexible as well.

Crounini

It's clever bu how do you do when you must routing more than 7/8 networks ?? You need necessarily a major hardware layer… (machine with 7/8 network devices...)

Supermule

No….I use VLAN's.... and only have 3 physical interfaces...

phatty

With the watchguard sitting on my desk I attempted to go ahead and update again today. This time I manually retrieved pfSense-2.1.4-RELEASE-4g-i386-nanobsd-upgrade.img.gz from the Bluegrass Mirror. This time update applied and system rebooted properly. Not sure if it was a problem with the auto updater selecting the wrong update, or if they fixed something with the files. I noticed the time stamp on the Bluegrass Mirror where I downloaded the update from was newer than when I had attempted the upgrade last.

stephenw10

@bennyc:

Making no progress here with that darn CF. Any hints on how to safely mount that BSD partition in windows?

The easiest way to do this, by far, is to do it from a BSD install. If you don't have a BSD machine then run a FreeBSD live CD like GhostBSD.

Steve

kpa

@stephenw10:

@bennyc:

Making no progress here with that darn CF. Any hints on how to safely mount that BSD partition in windows?

The easiest way to do this, by far, is to do it from a BSD install. If you don't have a BSD machine then run a FreeBSD live CD like GhostBSD.

Steve

Virtual machine with FreeBSD installed and connect the disk to it with an USB to SATA adapter, just about any USB enclosure will do.

phatty

Not that I'm surprised, but I thought it was worth reporting I'm now 2 for 2 on upgrading my Watchguard X750's using the manual update method.

pvoigt

@phatty:

Not that I'm surprised, but I thought it was worth reporting I'm now 2 for 2 on upgrading my Watchguard X750's using the manual update method.

Thank you for your almost realtime feedback on your update attempts. I would like to clarify one point before I am going to give my Alix backup machine a try: What do you mean by "manual update method"?
1.) From the webGUI: "System -> Firmware -> Manual Update" or 2.) From the console menu: "13) Upgrade from console"?

I have never used method 1.) before. Instead I used method 2.) or "Auto Update" from the webGUI.

I strongly suppose that update method does not matter as long as all selectable mirrors carry the same update files. Can you confirm the BluegrassNet mirror having had a more recent time stamp compared to the auto-selected mirror?

Regards,
Peter

bennyc

@phatty:

for a very short second, I was able to wait and eventually see the boot options for which slice to boot off of…

Thank you for that. While downloading the GhostBSD iso (also tnx at Steve for the hint), took a shot at this and after a couple of tries I got to boot slice 2. Now I have again console access, much better ;D

I will repeat upgrade attempt tomorrow evening. Are there logs kept on the system about past upgrade attempt? I'm curious to know where it went wrong, or what went wrong… (there's always something to learn)

bennyc

It is repeatable. I tried twice upgrading again, with exact same outcome as described here below.

Booted in 2.1.3, nano 2g image. Went to 'auto-update'

A new version is now available

Current version: 2.1.3-RELEASE
  NanoBSD Size : 2g
       Built On: Thu May 01 15:52:17 EDT 2014
    New version: 2.1.4-RELEASE

  Update source: https://updates.pfsense.org/_updaters

-> invoke auto upgrade

  Auto Update Download Status
----------------------------------------------------
  Current Version : 2.1.3-RELEASE
  Latest Version  : 2.1.4-RELEASE
  File size       : 76070249
  Downloaded      : 24739580
  Percent         : 33%
----------------------------------------------------

File size reflects here : https://updates.pfsense.org/_updaters/

latest-nanobsd-2g.img.gz                           25-Jun-2014 21:26            76070249

Seems to get the correct image in my case…

I tried following in console: this is partial output:


Broadcast Message from root@pfsense.localdomain
        (no tty) at 11:06 CEST...

Installing /root/latest.tgz.

Broadcast Message from root@pfsense.localdomain
        (no tty) at 11:13 CEST...

NanoBSD Firmware upgrade is complete.  Rebooting in 10 seconds.

*** FINAL System shutdown message from root@pfsense.localdomain ***

System going down IMMEDIATELY

pfSense is now shutting down ...

Jun 28 11:13:40 ovpns3: link state changed to DOWN
lighttpd[53693]:ovpns1: link state changed to DOWN
 (server.c.1558) server stopped pflog0: promiscuous mode enabled
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
Bump sched buckets to 256 (was 0)
pflog0: promiscuous mode disabled
Waiting (max 60 seconds) for system process `vnlru' to stop...done
Waiting (max 60 seconds) for system process `bufdaemon' to stop...done
Waiting (max 60 seconds) for system process `syncer' to stop...
Syncing disks, vnodes remaining...0 0 done
All buffers synced.
Uptime: 2h5m18s
usbus0: Controller shutdown
uhub0: at usbus0, port 1, addr 1 (disconnected)
usbus0: Controller shutdown complete
usbus1: Controller shutdown
uhub1: at usbus1, port 1, addr 1 (disconnected)
usbus1: Controller shutdown complete
usbus2: Controller shutdown
uhub2: at usbus2, port 1, addr 1 (disconnected)
usbus2: Controller shutdown complete
usbus3: Controller shutdown
uhub3: at usbus3, port 1, addr 1 (disconnected)
usbus3: Controller shutdown complete
usbus4: Controller shutdown
uhub4: at usbus4, port 1, addr 1 (disconnected)
usbus4: Controller shutdown complete
Rebooting...
þþþþÿ
1  pfSense
2  pfSense

F6 PXE
Boot:  1

After that, we were back at a dead box.
Interestingly, I never got the (very brief) serial console again until I cycled the box.
Could it be there is something wrong with this 2g image? Or am I missing something here?

For a small version upgrade, this is requiring way more attention then anticipated. ::)

stephenw10

Look at the update files here:
http://files.bgn.pfsense.org/mirror/updates/

Now look here:
https://updates.pfsense.org/_updaters/

Notice that the standard Nano upodate images are a different size to the VGA updates but at the _upaters URL they are the same (except the 512MB image). It seems likely they are all the VGA images.
Edit: Finally mangaed to open the files on Android. ::) Both have the same MD5 and both are decsribed as the VGA update in the MD5 file.

Go the manual update method by downloadibng the update file manually from the mirrors and uploading that to your box.

Steve