Bug 954547

Summary:	no partitions found on only HD /dev/sda
Product:	[openSUSE] openSUSE Tumbleweed	Reporter:	Felix Miata <mrmazda>
Component:	Installation	Assignee:	YaST Team <yast-internal>
Status:	RESOLVED WONTFIX	QA Contact:	Jiri Srain <jsrain>
Severity:	Major
Priority:	P5 - None	CC:	ancor, aschnell
Version:	Current
Target Milestone:	---
Hardware:	All
OS:	Other
URL:	https://trello.com/c/p8bVx55J
Whiteboard:
Found By:	---	Services Priority:
Business Priority:		Blocker:	---
Marketing QA Status:	---	IT Deployment:	---
Attachments:	two y2logs tgz sets in .zip from failing TW installation attempts on 32bit host gx280 y2logs.tgz from mirrors.us.kernel.org installation attempt .tgz of logs copied to USB from broken TW 20160307 attempt to 64 bit host big31 y2logs from i386 host gx280 tw20160404 installation attempt

Description Felix Miata 2015-11-10 23:43:32 UTC

Created attachment 655453 [details]
two y2logs tgz sets in .zip from failing TW installation attempts on 32bit host gx280

Attached are two sets of logs. The second is from a final try this evening using mirrors.us.kernel.org as source instead of download.opensuse.org. The latter set includes my mistake in not first ensuring en_US (which was Afrikaans initially) before clicking Next on license screen, then clicking Back to fix at first offered opportunity.

Mailing list thread:
http://lists.opensuse.org/yast-devel/2015-11/msg00066.html

This happened multiple times, then I wrote the mailing list explaining, then I tried again, achieving normal success. Later I tried again, and repeated the same failure, several times. Then trying yet again to get better logs and more detailed bug description and reproduction steps, success, 4 times. Last I tried the alternate installation source, which both succeeded and failed, then filed this.

To reproduce:
1-initialize static IP TW HTTP installation using Grub to load linux and initrd previously downloaded from mirrors (6 Nov. version)
2-click Next on license screen
3-click Next on Installation Options screen
4a-click create partition setup on suggested partitioning screen (which announces no automatic proposal possible), or
4b-click expert partitioner on suggested partitioning screen, next, then custom partitioning (for experts), next

Actual behavior:
1-(if 4a) screen offering custom partitioning for experts is (quickly) skipped
2-expert partitioner screen (quickly) appears showing entire /dev/sda is unpartitioned
3-fdisk -l on vtty2 shows all 31 existing partitions on /dev/sda

Expected behavior:
1-screen offering expert mode appears that when clicked results in
2-expert partitioner screen appearing,
3-showing all existing partitions on /dev/sda

Comment 1 Felix Miata 2015-11-10 23:56:08 UTC

Comment on attachment 655453 [details]
two y2logs tgz sets in .zip from failing TW installation attempts on 32bit host gx280

attachment 655453 [details] is the first y2logs.tgz set that I had intended to be in a combined .zip attachment

Comment 2 Felix Miata 2015-11-10 23:57:45 UTC

Created attachment 655455 [details]
y2logs.tgz from mirrors.us.kernel.org installation attempt

Comment 3 Ancor Gonzalez Sosa 2015-11-11 10:07:35 UTC

Added to YaST Team Scrum queue for the issue to be prioritized with the other tasks.

Comment 4 Felix Miata 2016-03-18 05:55:42 UTC

Created attachment 669584 [details]
.tgz of logs copied to USB from broken TW 20160307 attempt to 64 bit host big31

This just happened again slightly differently with tw 20160307. The create partition setup option existed and was entered, but otherwise was same as comment 0 in there being no partitions showing on sda, even though on tty2 fdisk -l shows 29 partitions.

Comment 5 Stefan Hundhammer 2016-04-05 13:30:28 UTC

Something very weird is going on here. From the y2log:

2015-11-10 11:17:06 <1> gx280(3206) [libstorage] SystemCmd.cc(execute):134 SystemCmd Executing:
"/usr/sbin/parted  -s  '/dev/sda' unit cyl print unit s print"

Model: ATA GB0500C4413 (scsi)
Disk /dev/sda: 60801cyl
Sector size (logical/physical): 512B/512B
BIOS cylinder,head,sector geometry: 60801,255,63.  Each cylinder is 8225kB.
Partition Table: msdos
Disk Flags: 

Number  Start     End       Size      Type      File system     Flags
 1      0cyl      4cyl      4cyl      primary   fat16           type=06
 2      5cyl      35cyl     31cyl     primary   fat32           hidden, type=1b
 3      36cyl     86cyl     51cyl     primary   ext2            boot, type=83
 4      87cyl     60800cyl  60714cyl  extended                  type=05
 5      87cyl     348cyl    261cyl    logical   linux-swap(v1)  type=82
 6      349cyl    349cyl    0cyl      logical                   hidden, type=11
 7      350cyl    381cyl    31cyl     logical   fat16           hidden, type=16
 8      382cyl    1095cyl   713cyl    logical   ntfs            type=07
 9      1096cyl   1707cyl   611cyl    logical   ext3            type=83
10      1708cyl   2319cyl   611cyl    logical   ext3            type=83
11      2320cyl   2931cyl   611cyl    logical   ext3            type=83
12      2932cyl   3543cyl   611cyl    logical   ext3            type=83
13      3544cyl   4461cyl   917cyl    logical   ext3            type=83
14      4462cyl   4722cyl   260cyl    logical   ext3            type=83
15      4723cyl   14921cyl  10198cyl  logical   ext3            type=83
16      14922cyl  15635cyl  713cyl    logical   ext3            type=83
17      15636cyl  16349cyl  713cyl    logical   ext3            type=83
18      16350cyl  17063cyl  713cyl    logical   ext3            type=83
19      17064cyl  17777cyl  713cyl    logical   ext3            type=83
20      17778cyl  18491cyl  713cyl    logical   ext3            type=83
21      18492cyl  19205cyl  713cyl    logical   ext3            type=83
22      19206cyl  19919cyl  713cyl    logical   ext3            type=83
23      19920cyl  20633cyl  713cyl    logical   ext3            type=83
24      20634cyl  21347cyl  713cyl    logical   ext3            type=83
25      21348cyl  22061cyl  713cyl    logical                   type=83
26      22062cyl  22775cyl  713cyl    logical                   type=83
27      22776cyl  23489cyl  713cyl    logical   ext3            type=83
28      59715cyl  60020cyl  305cyl    logical   ntfs            type=07
29      60021cyl  60734cyl  713cyl    logical   ntfs            type=d7
30      60735cyl  60799cyl  64cyl     logical   fat32           hidden, lba, type=1c
31      60800cyl  60800cyl  0cyl      logical                   hidden, type=11

Model: ATA GB0500C4413 (scsi)
Disk /dev/sda: 976773168s
Sector size (logical/physical): 512B/512B
Partition Table: msdos
Disk Flags: 

Number  Start       End         Size        Type      File system     Flags
 1      63s         80324s      80262s      primary   fat16           type=06
 2      80325s      578339s     498015s     primary   fat32           hidden, type=1b
 3      578340s     1397654s    819315s     primary   ext2            boot, type=83


The same disk shows up twice, once with 31 partitions, once with only 3. I wonder what could possibly get this inconsistence?


And the supplied disk_sda.info file shows the same disk without any partitions:

<?xml version="1.0"?>
<disk>
  <name>sda</name>
  <device>/dev/sda</device>
  <size_k>488386584</size_k>
  <major>8</major>
  <minor>0</minor>
  <readonly>true</readonly>
  <range>256</range>
  <geometry>
    <cylinders>60801</cylinders>
    <heads>255</heads>
    <sectors>63</sectors>
  </geometry>
  <label>msdos</label>
  <max_primary>4</max_primary>
  <ext_possible>true</ext_possible>
  <max_logical>255</max_logical>
  <udev_path>pci-0000:00:1f.2-ata-1.0</udev_path>
  <udev_id>ata-GB0500C4413_9QG0G0WS</udev_id>
  <udev_id>scsi-0ATA_GB0500C4413_9QG0G0WS</udev_id>
  <udev_id>scsi-1ATA_GB0500C4413_9QG0G0WS</udev_id>
  <udev_id>scsi-SATA_GB0500C4413_9QG0G0WS</udev_id>
  <transport>ATA</transport>
</disk>

There are also numerous errors "failed to get majorminor" for each of those partitions in the y2log, from /dev/sda1 to /dev/sda31.


I fear there is something wrong or inconsistent in your partition table. You can make the YaST expert partitioner create a brand-new one from scratch, but that will mean that all partitions are deleted in the process. Not sure if you'd want that.

I fear there is nothing we can do from remote. Sorry.

Comment 6 Lukas Ocilka 2016-04-05 13:35:31 UTC

Felix, you might want to reopen, but it's definitely nothing that Yast
can help with.

Either your GB0500C4413 HP 500-GB 1.5G 7.2K 3.5 SATA HDD is about to die,
or it's a Kernel problem, or ... but definitely nothing the Installer
can fix itself.

Comment 7 Felix Miata 2016-04-06 03:20:46 UTC

Created attachment 671774 [details]
y2logs from i386 host gx280 tw20160404 installation attempt

Today I initialized installation on 3 different Intel CPU machines, two using the 64 bit TW 20160404 installation set, 1 using 32 bit TW set. The 64s, one comment 4 host big31 (which had replicated the problem), the other host big41, both behaved as expected. The 32, comment 0 host gx280, misbehaved essentially the same as described in comment 0, no partitions found in the GUI, but expected fdisk -l output on tty2. Before running save_y2logs, I saved output of fdisk -l there, as well as a log from the partitioner I normally use, which provides a lot more detail about the existing partitioning, and what lives on the partitions, than any of the FOSS partitioning tools I've ever used.

After I got done, Ludwig's announcement of TW 20160405 arrived, so I booted 13.1, changed the partition type on /dev/sda29 from 0xD7 to 0x07 and started another installation. Its GUI found all existing partitions. At this point I booted back into 13.1, changed sda29 back from 0x07 to 0xD7, initiated another installation, and again the GUI found all existing partitions. This is wierd! If and when I see this happen on some other machine will probably be soon enough to revisit and maybe reopen this.