iSCSI root based servers appear to fail to boot completely

[James Page]

During alpha-3 ISO testing, iSCSI root volume based installs installed OK, but then failed to boot subsequently.

Very limited output to go on; DHCP requests seem to occur OK but sporadic output from initramfs.

I was able to get one test to boot by repeatedly rebooting the server.

[James Page (james-page)]

This occurs on amd64 and i386 images.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[James Page (james-page)]

OK so digging into this a bit more - the servers do actually boot; but from the console its not obvious in that it looks like the server has not completed bootup - but I can SSH to them all.

[Ubuntu QA Website (ubuntuqa)]

This bug has been reported on the Ubuntu ISO testing tracker.

A list of all reports related to this bug can be found here:
http://iso.qa.ubuntu.com/qatracker/reports/bugs/1028458

[Patrick Domack (patrickdk)]

I see my nic, but no ip set:
2: eth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop qlen 1000
    link/ether xx:xx:xx:xx:xx:xx brd ff:ff:ff:ff:ff:ff

I am not seeing any dhcp requests being done on the network

on all of my installs I get the following output:

ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
ipconfig: no devices to configure
/scripts/local-top/iscsi: .: line 505: can't open '/run/net-*.conf'
Gave up waiting for root device. Common problems:
.
.
.
(initramfs)

[Patrick Domack (patrickdk)]

Ok, it seems dhcp is missing from initrd. If I assign an ip manually, and exit initramfs propt, iscsi root gets mounted, and it boots fine.

I can't locate any dhcp client in the initrd image though.

[Patrick Domack (patrickdk)]

Ignore my last comment.

Seems running /scripts/local-top/iscsi, resets everything properly, and gets boot going again.

Possible this script is running before the kernel nic modules are loaded?

[James Page (james-page)]

Snapshot of one 'hanging' server that has actually booted AFAICT

[Ubuntu Foundations Team Bug Bot (crichton)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. It seems that your bug report is not filed about a specific source package though, rather it is just filed against Ubuntu in general. It is important that bug reports be filed about source packages so that people interested in the package can find the bugs about it. You can find some hints about determining what package your bug might be about at https://wiki.ubuntu.com/Bugs/FindRightPackage. You might also ask for help in the #ubuntu-bugs irc channel on Freenode.

To change the source package that this bug is filed about visit https://bugs.launchpad.net/ubuntu/+bug/1028458/+editstatus and add the package name in the text box next to the word Package.

[This is an automated message. I apologize if it reached you inappropriately; please just reply to this message indicating so.]

[Steve Langasek (vorlon)]

Hi James,

What's the console pointed at on this iscsi-booted system, and what's the kernel commandline look like? If the console is a graphical one, can you check which VT this fsck message is on?

[Steve Langasek (vorlon)]

More info needed from the submitter (James Page).

[James Page (james-page)]

I'll try to pull this information from an installed VM.

[James Page (james-page)]

Steve

The kernel command line is as installed by the installed for server.

I did notice that:

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash"

was set - uncommenting this did make things a little more verbose and I did get a login prompt.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Steve Langasek (vorlon)]

Assigning this to plymouth.

James, can you answer my other question, about which VT the fsck message shows up on when booting with 'splash'?

[James Page (james-page)]

I've struggled to reproduce this issue with the beta-1 images; it does seem to occur infrequently still (I get the Ubuntu 12.10 splash and never a login prompt) but more often than not it boots OK.

Steve - I think the VT I grabbed that screenshot off was tty1

[James Page (james-page)]

Hmm - I think this might be related:

Sep 4 14:04:50 ubuntu kernel: [ 5.120173] Console: switching to colour dummy device 80x25
Sep 4 14:04:50 ubuntu kernel: [ 5.121806] [drm:cirrus_vram_init] *ERROR* can't reserve VRAM
Sep 4 14:04:50 ubuntu kernel: [ 5.121809] cirrus 0000:00:02.0: Fatal error during GPU init: -6
Sep 4 14:04:50 ubuntu kernel: [ 5.121811] Trying to free nonexistent resource <00000000febf0000-00000000febf0fff>
Sep 4 14:04:50 ubuntu kernel: [ 5.121813] Trying to free nonexistent resource <00000000fc000000-00000000fc3fffff>

[Steve Langasek (vorlon)]

> Steve - I think the VT I grabbed that screenshot off was tty1

Well, it shouldn't have been, so I'd rather like to have confirmation there...

> Sep 4 14:04:50 ubuntu kernel: [ 5.121806] [drm:cirrus_vram_init] *ERROR* can't reserve VRAM

This is bug #1038055.

[Steve Langasek (vorlon)]

According to James, this problem is not reproducible if using the vga driver under kvm instead of the cirrus driver. So marking as a duplicate of the kernel bug.