Update to refocus.
I think I need to refocus the subject.
Original issue: Systems are booting on BSOD after image deployment after doing an upgrade on FOG.
Facts: Images have not been changed. Hardware has not been changed. Hardware has not been reconfigured. FOG was updated.
More images are being discovered to boot systems into BSOD but not all.
Discovered: Missing fog user in MySQL user table. fog user still presents in passwd file. Password for storage same as password .fogsettings.
Assumptions: Possible corrupted images. Missing permissions, wrong access through MySQL.
Another image was found doing BSOD today. This is an old image about 3 months old.
Still collecting more data and a debug.
If MySQL users were messed up, your web interface wouldn’t work. FOG doesn’t setup a ‘fog’ user for MySQL. Typically it uses the root user without password to access MySQL if you’ve not manually defined a user/password in MySQL.
I doubt it’s permissions. It could be a corrupted image - but if that’s the case, it’s likely that the server HDD is dying or underwent a sudden power loss.
What’s stopping you from making a new image and trying that?
I don’t fully understand.
The MySQL fog.users table only manages FOG WEB GUI logins. Not have a ‘fog’ user here is not the end of the world unless you are trying to login to the GUI with a user named FOG. It has no reference for FTP/Linux logins. (Your environment already seems to have this working as well as – from what I can tell – uploads still work?).
The fact that images are suddenly “bsoding” might be something that FOG has done, but this seems unlikely and probably something many other people would’ve brought up by now. Yes, there’s enough users running trunk/dev versions of fog that I feel pretty confident there’d have been at least 2 - 10 other posts saying the same thing by now. This would then lead me to believe something is happening on the server. Maybe a drive is dying? Maybe the network is corrupting the traffic as it’s being placed on the system? I don’t know what it is, but based on the “randomicity” of the BSOD at this point, it seems something to think about.