[Nova][CentOS] extrem slow first boot Centos 8.1 with failures

Nicolas Ghirlanda nicolas.ghirlanda at everyware.ch
Wed May 27 14:47:13 UTC 2020


Hello all,


I created a centos image with a pretty similar kickstart config like 
from that link

https://github.com/CentOS/sig-cloud-instance-build/pull/159/commits/2c542bb2b1bc54c007fbf57a5da0a3213ce73fbb


Packer is building the image fine, also the post installation tasks 
after a reboot in qemu (which is running fine and fast), so everything 
looks good so far.


Then upload to openstack and on first boot, the instance is booting fine 
until that point of boot:


[[0;32m  OK  [0m] Started D-Bus System Message Bus.
[[0;32m  OK  [0m] Started Hardware RNG Entropy Gatherer Daemon.
          Starting OpenSSH rsa Server Key Generation...
          Starting System Security Services Daemon...
[    9.913211] piix4_smbus 0000:00:01.3: SMBus Host Controller at 0x700, revision 0
[    9.988763] input: PC Speaker as /devices/platform/pcspkr/input/input6
[   10.137969] sd 0:0:0:0: Attached scsi generic sg0 type 0
[   10.332234] cirrus 0000:00:02.0: vgaarb: deactivate vga console
[   10.419654] Console: switching to colour dummy device 80x25
[   10.421488] [TTM] Zone  kernel: Available graphics memory: 2017692 kiB
[   10.422946] [TTM] Initializing pool allocator
[   10.424036] [TTM] Initializing DMA pool allocator
[   10.425630] [drm] fb mappable at 0xFC000000
[   10.426612] [drm] vram aper at 0xFC000000
[   10.427436] [drm] size 33554432
[   10.428108] [drm] fb depth is 16
[   10.428884] [drm]    pitch is 2048
[   10.449202] fbcon: cirrusdrmfb (fb0) is primary device
[   10.465230] Console: switching to colour frame buffer device 128x48
[   10.499394] cirrus 0000:00:02.0: fb0: cirrusdrmfb frame buffer device
[   10.524033] [drm] Initialized cirrus 1.0.0 20110418 for 0000:00:02.0 on minor 0


On the console there is no further update from here.

In the boot log it slowly progresses, but with Failures of starting services

e.g.


          Starting RPC Bind...
          Starting Security Auditing Service...
[    9.500362] audit: type=1400 audit(1590589814.846:4): avc:  denied  { read } for  pid=784 comm="auditd" name="group" dev="sda1" ino=259654 scontext=system_u:system_r:auditd_t:s0 tcontext=system_u:object_r:unlabeled_t:s0 tclass=file permissive=0
[    9.517988] audit: type=1400 audit(1590589814.863:5): avc:  denied  { read } for  pid=783 comm="rpcbind" name="passwd" dev="sda1" ino=259634 scontext=system_u:system_r:rpcbind_t:s0 tcontext=system_u:object_r:unlabeled_t:s0 tclass=file permissive=0
[[0;1;31mFAILED[0m] Failed to start RPC Bind.
See 'systemctl status rpcbind.service' for details.

or

[[0;32m  OK  [0m] Started OpenSSH ecdsa Server Key Generation.
[[0;32m  OK  [0m] Started GSSAPI Proxy Daemon.
[[0;32m  OK  [0m] Started OpenSSH ed25519 Server Key Generation.
[[0;32m  OK  [0m] Started OpenSSH rsa Server Key Generation.
[[0;1;31mFAILED[0m] Failed to start Authorization Manager.
See 'systemctl status polkit.service' for details.
[[0;1;33mDEPEND[0m] Dependency failed for Dynamic System Tuning Daemon.
[[0;1;31mFAILED[0m] Failed to start NTP client/server.
See 'systemctl status chronyd.service' for details.


Also cloud-init is not able to assign an ip.

[[0;32m  OK  [0m] Started D-Bus System Message Bus.
          Starting Network Manager...
[  550.962041] cloud-init[1070]: Cloud-init v. 18.5 running 'init' at Wed, 27 May 2020 14:37:46 +0000. Up 460.76 seconds.
[  550.964516] cloud-init[1070]: ci-info: +++++++++++++++++++++++++++Net device info++++++++++++++++++++++++++++
[  550.967275] cloud-init[1070]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+
[  550.970075] cloud-init[1070]: ci-info: | Device |   Up  |  Address  |    Mask   | Scope |     Hw-Address    |
[  550.972524] cloud-init[1070]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+
[  550.975221] cloud-init[1070]: ci-info: |  ens3  | False |     .     |     .     |   .   | fa:16:3e:a3:7d:d5 |
[  550.977970] cloud-init[1070]: ci-info: |   lo   |  True | 127.0.0.1 | 255.0.0.0 |  host |         .         |
[  550.980623] cloud-init[1070]: ci-info: |   lo   |  True |  ::1/128  |     .     |  host |         .         |
[  550.983300] cloud-init[1070]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+
[  550.986164] cloud-init[1070]: ci-info: +++++++++++++++++++Route IPv6 info+++++++++++++++++++
[  550.988469] cloud-init[1070]: ci-info: +-------+-------------+---------+-----------+-------+
[  550.991092] cloud-init[1070]: ci-info: | Route | Destination | Gateway | Interface | Flags |
[  550.993576] cloud-init[1070]: ci-info: +-------+-------------+---------+-----------+-------+
[  550.996326] cloud-init[1070]: ci-info: +-------+-------------+---------+-----------+-------+



After that time, the system is up, but no hostname is set (due to 
cloud-init issues).


If I now reboot, everything is working fine, cloud-init is working etc.


If I set selinux from enforcing to permissive in the kickstart config, 
the issues are not showing up, but I need selinux to stay enforcing.


Anyone with the same issues or any hints appreciated.


regards


Nicolas









-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20200527/b4b64f06/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2818 bytes
Desc: not available
URL: <http://lists.openstack.org/pipermail/openstack-discuss/attachments/20200527/b4b64f06/attachment.bin>


More information about the openstack-discuss mailing list