Boot failure - could not mount packagefs

[fishpond]

Depite any boot options (safe, SMP, etc.), on my core I7 quadcore, I'm going to KDL on the disk icon. This includes USB and SATA disk boots. The boot volume itself is found, but packagefs is not. The same USB stick boots fine on other computers. Variations of BIOS settings regarding USB (EHCI handoff, leagcy OS) do not help either. Pre-packagefs builds do boot fine. On subsequent boots on other comupters, no traces can be found in the syslog. Have tried various nightlies to exclude scheduler changes. Please indicate if there is any way I can provide more information.

[anevilyak]

Two things would be of interest: 1) a picture of the KDL screen in question, and 2) if you're able to reboot from the KDL prompt, entering the boot loader on the subsequent boot should allow you to view the previous (failed) boot's in-memory syslog, and also save it to a fat32 formatted USB stick if you have one available/inserted. If that can be acquired, attaching it to this ticket would be a starting point.

Assigning to system/kernel until a better candidate can be found.

[fishpond]

KDL Screenshot

[fishpond]

Thanks for the tips. I did not manage to get a log since my mainboard does not have PS/2 connectors and USB keyboards are not yet up at the time of the KDL. During different reboots, I started fiddling BIOS settings again and managed to get a successful USB and SATA disk boot. The disk controller can be set to IDE emulation, AHCI or RAID mode. I had set it to RAID, since AHCI would not let me boot into pre-packagefs Haiku installed on a harddisk and IDE mode would prevent command queueing. RAID mode btw does not actually require any disks being put into a RAID group, the controller will just give itself a different PCI-ID and be ready to accept some RAID-specific commands.

It seems that now, IDE and AHCI modes are the possible options. What strikes me is that setting controller mode to RAID does effectively prevent USB boots. Still, I'm happy to be back on native Haiku and start to adapt to the packaged world order. In absence of a real blocking problem and insufficient evidence of the problem encountered, I think it'd be perfect to close the ticket and remember the workaround for any other users with ICH10 chipsets.

Thanks again.

[kallisti5]

Is this still an issue on a recent nightly image?

[fishpond]

Problem no longer occurs - just found this ticket back.

[anevilyak]

Thanks for the update!