Upgrade from 20.1.9_1 to 20.7 failed - no sig file found

[jkemp]

Can someone help me with a failed system update. Now it's stuck and won't upgrade. No connectivity for clients but can ping from shell.

When I try and update from the OPNsense web client I get a response of:
Firmware status check was aborted internally. Please try again.

***GOT REQUEST TO UPGRADE***
Fetching packages-20.7-OpenSSL-amd64.tar: .. failed, no signature found
***DONE***

When I use option 12 to update from console it states nothing found. I entered the 20.7 major update response at at prompt but it can not find the sig file?

What can I do to save this system, any help would greatly be appreciated.

James

[pouakai]

Yes, been through this too.  There is also another topic here https://forum.opnsense.org/index.php?topic=18361.0 which sounds like the same issue.

In my case the upgrade continued and FreeBSD 12 was successfully installed, but Opnsense 20.7 failed to install (bad sig) (Opnsense 20.1 remained).  This, of course, left quite a mess.

I did fix it by using the opnsense-bootstrap.sh script on Github, though I had to change FreeBSD 11 --> 12 and 20.1 --> 20.7 in the script.  I also had to play with the routing to get the files to download through the LAN, since WAN was borked.  Then I uploaded a backup to get it all working again.

It probably would have been easier to just re-install.

 

[franco]

There is only one single reason for "no signature found" error: the update files are not on the selected mirror.


Cheers,
Franco

[alex_rhys-hurn]

@Franco, yes you are right, and the reason, if you look at my other thread, is that the Repo file in /usr/local/etc/pkg/repos/OPNsense.conf gets borked.

Its requesting OPNsense 20.1 with FreeBSD (HBSD) version 12.... Which doesnt exist.

Why is this happening? (EDIT: Genuine Technical Curiosity question, not frustration)

Alex

See this post: https://forum.opnsense.org/index.php?topic=18361.0

[franco]

Could be, but not enough contexts from the original report.

If that should be the case the same goes here: please post your /var/cache/opnsense-update/.upgrade.log


Cheers,
Franco

[jkemp]

Changing the OPNsense.conf file fixed it. Thank you for the help.

James

[franco]

First time upgrade didn't work, but second time "something" did? Makes no sense to me so log would still be helpful.


Cheers,
Franco

[pouakai]

Well, this may not much use, but here goes.  This is the .upgrade.log file that I have after running opnsense-bootstrap.sh.  Looking at the creation time, it may have been created at the first upgrade attempt - I just can't remember when I started exactly.

The file cuts off prematurely.  I don't why; it's the file that I have.

Just adding:  My usual mirror (Australia) gave the bad sig message and stopped immediately, and I then tried the San Francisco mirror, guessing it was a server issue.  Same thing.  I then tried again, and it didn't give an error; the upgrade continued, but the upgrade failed.

On upgrade, I saw Opnsense 20.7 being 'installed', and the installer said something along with  'Nothing to do!'. The installer moved onto FreeBSD 12 (which did install), then proceeded to reboot. 

Let me know if you need any more information.

[franco]

> [5/125] Extracting python37-3.7.8_1: .....

Depending on the disk used the Python installation can take quite a lot of time (more than 30 minutes). Did you force a reboot at some point in the upgrade process?


Cheers,
Franco

[pouakai]

Yes, after the web interface said the was system rebooting, the web interface did not change for a long, long time. I tried to ping the box, but it would not respond.  In the end I forced it off by holding the power button.  Took the box away for a closer inspection...

There was no console monitor on the box, so I had no idea what it was doing.  I had an SSH session open, but it closed when the web interface said rebooting.

Box is a Jetway NUC with 3160 processor.  Disk is a SSD SATA with ZFS.

Hmmm... I should have waited longer?

[mimugmail]

Hmmm... I should have waited longer?

Yes .. python37-3.7.8_1 and ruby can take ages

[fruit]

My HP T610 took a little under two and a half half hours to complete the upgrade - no complaints, just an observation

[pouakai]

O.K. Thanks for the replies.

The Jetway does not have a HDD Activity LED, so it's hard to know it's alive.  Next time, I will leave a console connected to see what is happening