Opened 7 years ago

Closed 7 years ago

#911 closed defect (wontfix)

bootloader install failed on install to hard drive

Reported by: mattjs Owned by: live-demo@…
Priority: normal Milestone:
Component: OSGeoLive Keywords: osgeo-live-5.5.iso; bootloader; install; grub; install dialog
Cc:

Description

"Booloader install failed" cannot be resolved or overridden by either:

  • "Choosing a different device to install the bootloader on:" if it fails one cannot --force the grub-install which works from terminal under regular Ubuntu Desktop OS) or ignore and "Cancel the installation." which doesn't work and simply retuns to the dialog box to prompt again.

Fortunately this appears to occur, as it should, after installing everything to disk so killing the processes/OS at this point leads to recovery (for me). [In my case, although this doesnt effect any of the above, I dual-boot to phys disk install using NeoSmart? EasyBCD autogrub (aka wingrub of old) and virtual boot in VMware using the phys disk install and a supergurb disk. I intall in VMware using phys disk aloowing access to the swap and root partitions only and in line with the above always install grub? to the partition itself e.g. /dev/sda4 swap and /dev/sda5 or sda6 say (Ubuntu or OSGeoLive). Writes to partition table/mbr always fail and never trash my Windows install. Writes to sda5 and sda6 with grub-install --force work and AutoNeoGrub?(wingrub) for phys boot and super_grub_disk for virtual boot then do the rest! :-)]

Attachments (1)

Cannot Continue without a bootloader.jpg (210.9 KB) - added by mattjs 7 years ago.
Cannot "Continue without a bootloader"

Download all attachments as: .zip

Change history (11)

comment:1 Changed 7 years ago by mattjs

I meant to say "Cannot install without a bootloader" then couldnt Cancel also afterwards.

Changed 7 years ago by mattjs

Cannot "Continue without a bootloader"

comment:2 Changed 7 years ago by mattjs

So the upshot is that someone needs to review this dialog box code and options and make sure that it allows "Continue install without a bootloader" and "Cancel..." and/or perhaps a "-force" install options or just ignore the error and "Continue..." as it should above. I presume that one or other of the boot options on the LiveCD will probably permit recovery anyway in which case not preventing continuation even for a novice users wouldn't then be a big deal (and they are likely to leave the CD in the tray anyway).

comment:3 Changed 7 years ago by wildintellect

Just to confirm this was using the 5.5 iso (or mini iso) and using the install link on the desktop? Did you try the install option from the boot menu?

comment:4 Changed 7 years ago by mattjs

Keywords: osgeo-live-5.5.iso added

osgeo-live-5.5.iso

comment:5 Changed 7 years ago by mattjs

ah, from the install link on the desktop of booted, running osgeo-live-5.5.iso CD in VMware Workstation 8.0.4 build-744019. (I cannot see that making any difference in this case but its just these sorts of bugs that make me feel all warm gooey and smart inside when I knowingly install with the rest of my hard disk partitions nicely locked away from any bogeys although I always backup before an install or a re-partition anyway. Also explains my over-elaboration of these unrelated issues above.) Maybe I shoutd volunteer: is it python code at this point (+what gui? not tcl/tkinter but somethign else?)

comment:6 Changed 7 years ago by wildintellect

We do not code this ourselves, it comes directly from whatever Remastersys puts on the disc. This method is subject to change as our build method is moving towards what the regular ubuntu disc does. The best thing you could do is test our 6.0 alpha builds and compare the 2 build methods, as we have an alpha for each.

http://live.osgeo.org/dev/build/

Alternate is the new build method.

comment:7 Changed 7 years ago by mattjs

I see. BTW: I had used the custom install method choce just before this: selecting my original swap and re-formatting and re-installing over an existing osgeo 5.0 root. It is almost always the custom options that have the bugs. Someone upstream will work it out and fix it no doubt and if you are using new install method for future releases then it likely wont be an issue. It was receoverable for me anyway it seems.

comment:8 Changed 7 years ago by mattjs

BTW: whats the policy for installing the suggested Ubuntu updates that come through? I am loath to do so unless I feel a specific need to for a bug fix or version upgrade as I fear that something from the upstream ubuntu tree might break the Custom OSGeo Install. Are my concerns legitimate here? Also the kernel sources are not included so VMware's custom tools will not install without but the vmware graphics and mouse kernel modules from the tree are already pre-included and whilst not the greatest and latest (e.g. for automatic VM screen/window resizing) they should I think suffice even for intensive GIS graphics. Next OSGeoLive release then next install will be my policy. I have an ugradable Ubuntu install and OSGeo4W so I am covered for rendundancy and persisting any particular application I choose to along with its data. Regards, Matthew

comment:9 in reply to:  8 Changed 7 years ago by wildintellect

Replying to mattjs:

BTW: whats the policy for installing the suggested Ubuntu updates that come through?

Please join the mailing list to discuss other questions that are not specific Issue Tickets. http://lists.osgeo.org/mailman/listinfo/live-demo

comment:10 Changed 7 years ago by wildintellect

Resolution: wontfix
Status: newclosed

Possibly resolved with switch to Dist mode instead of Backup in remastersys. Closing until testing with 6.0 release.

Note: See TracTickets for help on using tickets.