Newest Snapshot broke system

  • So I installed the newest 1.2.3RC2 Snapshot about 1hour ago, and now it seems to have broken the system.  I get many PHP startup errors, about being unable to initialize the module, and a fatal error, unable to call xmp_parser_create() in /etc/inc/

    My last snapshot update was from about 2 days ago, which has been working perfectly.  The pfsense console menu and local shell access seems to work ok, though I definitely have no LAN connectivity, and I am unsure about WAN, though I am guessing not.  I am wondering if there is anyway to restore my old firmware from 2days ago, using the local shell?  Do I have any options other than a full system restore?

    If my only option is a full system restore, what about my current configuration?  The only copies I have of it, are on the pfsense system.  Anyway to use or recover this?

  • Same here.

    I had saved my config though so Im up and going again…

    Good luck!

  • Just a update, I was unable to correct the issue using Shell Access from the console.  I then burned an .iso from the July 2nd snapshot.  I attempted using 'rescue config.xml', but this also did not work properly, and just produced errors.  In the end, I ended up having to do a fresh install, from the July 2nd snapshot.  Unfortunately I did not have an offsite config backup, so I have to start fresh.  Good lesson learned though, always test updates first on a none use vmware dummy setup, and always backup configs and such before updates.  What can you do.  It would still be interesting to know what happened with this snapshot,  I have heard from a few friends now as well, that snapshot updates today completely borked their system.

  • System died for me too, back to WRT54GL for now. I gonna test the nanobsd build later on when i have figured out howto make it run on a pc with a IDE/CF card convertor. I was hoping that systenm will run after flash it on a CF card and boot it up.

  • I had the same with this snapshot (not Upgrade):


    but this one works:


    both seems to be from 8th of July. The one which not works a few hours earlier.

    All is working, but in System/Packages it's not possible to insert any packages. Maybe the filesystem on CF is read only? Could be.

    Installation of AutoConfigBackup FAILED!

    Downloading package configuration file… failed!
    Installation aborted.
    Installation halted.


  • Same for me,
    Everything halted. Maybe I need to do a fresh install?

    Somewhat disappointed..

  • Samething happened to me last night on a production system, Ive had 3 hours sleep  :D

    Upgrades churn out a lot of errors and it looks like some sort of php modules symbols error if remember correctly, full installs of the same date seem to work fine with the exception of an error that shows up after system start up beeps.

    Starting captive portal…
    ipfw2 (+ipv6) initialized, divert loadable, nat loadable, rule-based forwarding enabled, default to accept, logging disabled

    Fatal error: Call to undefined function: mwxec() in /etc/inc/ on line 1085
    Starting CRON...
    Syncing packages


    EDIT: Using 1.2.3-RC2
    built on Wed Jul 8 19:40:50 EDT 2009
    FreeBSD 7.2-RELEASE-p2 i386

  • Update: I don't know why, but I have 2 pieces 4 Giga CF and one 2 Giga CF. With one of the 4 Giga CF I have the same problem as written here. the other works well and I write the image allways the same way. The one which don't works is an slc type.


  • @hansru:

    Same for me,
    Everything halted. Maybe I need to do a fresh install?

    Somewhat disappointed..

    I was somewhat disappointed at first also, but lets look at the big picture, these are release candidate snapshots.  Not full out tested stable releases.  If this is going on a production or 'must be in service' system, maybe the update should of been tested first via vmware?  or maybe you should just be using a known 'STABLE' release, instead of basically bleeding edge.  I was surprised though, that the live cd rescue config option did not work either.  Hopefully people had backups in place.  Lessons learned for everyone, should be common practice, backups & testing before putting into production!!

    Any another note, does anyone have word on what the issue was?  Has it been corrected in the most recent snapshots?

  • From the liveCD, one should be able to mount the hard drive and copy the config.xml even if the console "rescue config.xml" fails, as I did (and then I found my backup :P). Or was that a problem for some?

    A backup /conf/config.xml on an inserted thumbdrive is a great way to reinstall in case of calamity on a PC system. Does that work with CF as well (assuming you have an usb port)?

    Some of the *.inc scripts in /etc/inc of the failed update  were quite different from those present once I got back up. So I suspect there were changes under the hood that turned into trip wires…


  • "Any another note, does anyone have word on what the issue was?  Has it been corrected in the most recent snapshots?"

    Not sure, as I am sitting tight on the fresh working install.  As Slam hinted, there may have been a difference between updates and fresh installs, ie. a problem with the updates….


  • No, for me it is not fixed at the moment, I have updated last night with a recent snapshot and it breaks my system.

    Wait and take a look in the repository für 1.2x…....


  • Is it safe to upgrade?

  • Upgrade from what? 1.2 to 1.23 RC2??

  • yes :)

  • Just an update, it appears all is good now.  Just updated:

    New version:    Sun Jul 12 04:51:52 EDT 2009
    Current version: 1.2.3-RC2

    built on Sun Jul 12 01:54:58 EDT 2009

    No issues what so ever, or errors.  All seems good.  Wonder if the issues had to do with the full file system on the snapshot server??

  • Im back on pfSense  ;D

    Lucky me i have WRT54GL with Tomato as backup.

Log in to reply