Boot failure issues

[scottmc]

Use this as a tracking ticket for other related driver trac tickets ...

Prop #27 Boot failure issues

• All known bugs that cause boot failures on supported hardware to be resolved

Developer results

• must-have (nielx, pulkomandy, siarzhuk, anevilyak, 3dEyes, stippi, leavengood, mmadia, colin, zooey, ithamar, laplace, tqh, phoudoin, bonefish, humdinger, Wim, sikosis, brecht)
• only-if-ready (czeidler, jprostko, umccullough, jonas.kirilla, scottmc, axeld, mmlr, yourpalal, korli)

General poll results

[mmadia]

edit: removed comment that was meant for the ticket description.

[scottmc]

This is not a feature, so not required for Beta1, moving to R1

[luroh]

(In #9172) At a minimum, we need to know what image you downloaded (raw, iso, anyboot?) and how you are trying to boot from it (USB stick, CD, hard disk?). Describing how far booting gets also wouldn't hurt, either by words or by attaching a photo. For reference, see ReportingBugs.

[luroh]

(In #9169) Any chance of a grabbing a syslog?

[luroh]

(In #9174) Sounds like a bad image, dd or USB stick. Assuming you are on a *nix, try to re-download, dd, and run 'sync' before restarting or pulling the stick out, to make sure everything is flushed to the stick. If that doesn't work, try another USB stick if possible.

[luroh]

(In #9097) Any chance you could provide the version, method of booting and a syslog?

[luroh]

(In #9068) Any improvements lately? Could you please provide a syslog?

[luroh]

(In #8623) grinder: could you please give R1 Alpha 4.1 a try?

[waddlesplash]

(In #11202) I see no reason to think this is x86-specific. Also, can you please try a more recent nightly from ​http://download.haiku-os.org/ and see if it's fixed? And if it isn't, can you do binary searching to find which revision broke it?

Also, if you find a working version, the syslog and listdev would be quite useful.

[waddlesplash]

(In #11202) And my bad again -- this is really a duplicate.

[pulkomandy]

(In #2177) Not a strict "boot failure" as the system boots eventually, removing from #7665.

[pulkomandy]

(In #2923) Not a boot issue, removing from #7665.

[pulkomandy]

(In #3868) Not a boot failure...

[pulkomandy]

(In #4522) No boot failure, it is only slow.

[pulkomandy]

(In #4570) Slow reboot != boot failure.

[pulkomandy]

(In #5880) shutdown issue != boot issue.

[pulkomandy]

(In #6039) slow boot != boot failure.

[pulkomandy]

(In #7440) reboot issue != boot issue.

[pulkomandy]

(In #10055) slow boot != boot failure

[pulkomandy]

(In #10198) PXE specific issue, not relevant to the "boot failure" ticket.

[pulkomandy]

(In #10139) R1 target is "IBM PC compatible" and that doesn't include UEFI. Feel free to fix it anyway, but removing from #7665.

[pulkomandy]

(In #10864) mh... anything more we can do here? Not being able to boot when the FS is corrupt sounds ok to me, especially if checkfs can fix it.

[pulkomandy]

(In #11135) Not an issue for standard images, removing from #7665.

[pulkomandy]

(In #9195) Ok, that's fine. I'm changing the ticket to "enhancement" since it was never announced that we would support this.

You may try to build a boot image manually using mkdiskimage from linux: ​http://www.pendrivelinux.com/booting-linux-from-usb-zip-on-older-systems/

Then copying our raw partition to it and running makebootable on it. I don't know if this is enough to get usb-zip working, or if more tricks are needed.

An alternative is booting from a CD, or using a boot manager with USB support such as plop (​http://www.plop.at/en/bootmanagers.html) to boot Haiku from USB.

[pulkomandy]

(In #6811) gcc4-only issue, moving out of R1.

Does this still happen btw?

[waddlesplash]

(In #11684) According to your syslog, it appears package_daemon mounted everything and the video driver (intel_extreme) was just starting when the KDL occured. I couldn't find anything else useful.

[waddlesplash]

The list here has become *very* large and rather unwieldy to navigate. Perhaps we should add a boot-failure keyword to all still-open tickets attached to this ticket and close it?

[pulkomandy]

Well, I think we will never be able to close this completely, anyway? If we keep this open, R1 will never be released. Maybe it's time for a refresh of the "R1 feature poll" voting.

[waddlesplash]

Working on a case-by-case basis as we've been doing for beta1 sounds good to me -- let's see if we can get a consensus on that. I'll see about adding a keyword to the open tickets.

[waddlesplash]

OK, so, for whatever reason, I can't batch-modify add a "boot-failure" keyword -- it just silently fails. Going through and manually adding a keyword to 100 tickets does not sound fun.

[waddlesplash]

And it's now been done. Closing this as "fixed", as the boot failure tickets are now tracked using the boot-failure keyword, which you can see the open items of using a query.