[Bug 1623383] Re: Some restarts fail due to missing base devices

Stefan Bader stefan.bader at canonical.com
Fri Sep 23 13:36:39 UTC 2016


Adding the setup of the zVM guest that shows the problem in case that
helps to re-create the problem:

Memory: 4G, 8 vCPUs

Device   Subchan.  DevType CU Type Use  PIM PAM POM  CHPIDs
----------------------------------------------------------------------
0.0.e002 0.0.0000  1732/03 1731/03 yes  80  80  ff   20000000 00000000 
0.0.e102 0.0.0001  1732/03 1731/03 yes  80  80  ff   21000000 00000000 
0.0.0200 0.0.0002  3390/0c 3990/e9 yes  80  80  ff   ff000000 00000000 
0.0.0201 0.0.0003  3390/0c 3990/e9 yes  80  80  ff   ff000000 00000000 
0.0.0009 0.0.0004  0000/00 3215/00 yes  80  80  ff   ff000000 00000000 
0.0.000c 0.0.0005  0000/00 2540/00      80  80  ff   ff000000 00000000 
0.0.000d 0.0.0006  0000/00 2540/00      80  80  ff   ff000000 00000000 
0.0.000e 0.0.0007  0000/00 1403/00      80  80  ff   ff000000 00000000 
0.0.0600 0.0.0008  1732/01 1731/01 yes  80  80  ff   00000000 00000000 
0.0.0601 0.0.0009  1732/01 1731/01 yes  80  80  ff   00000000 00000000 
0.0.0602 0.0.000a  1732/01 1731/01 yes  80  80  ff   00000000 00000000

0.0.e002 and e102 are two zfcp adapter leading to one disk via two paths
each. The resulting 4 paths are grouped by multipath and make up a lvm2
VG. But nothing is mounted from there by default.


lsblk
NAME                                MAJ:MIN RM  SIZE RO TYPE  MOUNTPOINT
sdd                                   8:48   0   10G  0 disk  
`-36005076306ffd6b60000000000002a03 254:1    0   10G  0 mpath 
dasda                                94:0    0  6.9G  0 disk  
`-dasda1                             94:1    0  6.9G  0 part  /
sdb                                   8:16   0   10G  0 disk  
`-36005076306ffd6b60000000000002a03 254:1    0   10G  0 mpath 
dasdb                                94:4    0 20.6G  0 disk  
`-dasdb1                             94:5    0 20.6G  0 part  
  `-datavg-home                     254:0    0 20.6G  0 lvm   /home
sdc                                   8:32   0   10G  0 disk  
`-36005076306ffd6b60000000000002a03 254:1    0   10G  0 mpath 
sda                                   8:0    0   10G  0 disk  
`-36005076306ffd6b60000000000002a03 254:1    0   10G  0 mpath

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to udev in Ubuntu.
https://bugs.launchpad.net/bugs/1623383

Title:
  Some restarts fail due to missing base devices

Status in udev package in Ubuntu:
  Confirmed

Bug description:
  Arch: s390x
  Release: Yakkety / 16.10

  This happens on some (but not all) system starts with Yakkety. In
  Xenial (which is using the same 4.4 kernel version the Yakkety systems
  were using when the problem was first observed) this did not happen.
  The system (LPAR) this was seen first was an upgrade from Xenial but
  since then has been freshly installed with Yakkety. The same behaviour
  is seen on a zVM guest running Yakkety.

  The attached syslog shows a failed boot, followed by one that did work. Note the "Found device .*(sclp|encc00).*" messages in the good boot. Those are missing in the bad attempt and as a result networking and console fail to be usable. Also note, those boots were 4.8 kernels but we saw this with 4.4 kernels, too.
  This might be a systemd problem / race, I just filed it into udev for now as that better matches the not finding basic devices symptom.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/udev/+bug/1623383/+subscriptions



More information about the foundations-bugs mailing list