Dear Folks,
After a power surge and a spontaneous reboot, my 4-core F15 x86_64 machine has been unhappy. I have forced a selinux relabel, and a forcefsck.
Symptoms: 1. The graphical boot did not complete: it started Nagios, but cupsd failed to start, and gdm did not start up. 2. I cannot log in on the text consoles; I can type the username, but no prompt appears for the password. 3. Attempts to use sudo just hang. 4. I can log in by ssh as either myself or as root. 5. named is using "97.5%" CPU, though when I turn on querylogging, there are very few lookups. 6. Attempts to start cups result in complaints about dbus: # systemctl start cups.service Failed to get D-Bus connection: Failed to authenticate in time.
/var/log/messages says nothing of the matter.
Now systemctl is new to me, and I am not sure how to diagnose this. I would be grateful for any suggestions. Where should I look for further clues?
Okay, now I see that cupsd and dbus processes are present, but not accepting connections. And further: # systemctl kill cups.service Failed to get D-Bus connection: Failed to authenticate in time. # systemctl Failed to get D-Bus connection: Failed to authenticate in time.
I have setenforce 0, and the behaviour appears unchanged. rpm -V indicates no apparent problems with pam, systemd, cupd, or any of the dbus packages. Any suggestions most welcome.
On 28/09/11 16:47 +1000, Nick Urbanik wrote:
Dear Folks,
After a power surge and a spontaneous reboot, my 4-core F15 x86_64 machine has been unhappy. I have forced a selinux relabel, and a forcefsck.
Symptoms:
- The graphical boot did not complete: it started Nagios, but cupsd failed to start, and gdm did not start up.
- I cannot log in on the text consoles; I can type the username, but no prompt appears for the password.
- Attempts to use sudo just hang.
- I can log in by ssh as either myself or as root.
- named is using "97.5%" CPU, though when I turn on querylogging, there are very few lookups.
- Attempts to start cups result in complaints about dbus:
# systemctl start cups.service Failed to get D-Bus connection: Failed to authenticate in time.
/var/log/messages says nothing of the matter.
Now systemctl is new to me, and I am not sure how to diagnose this. I would be grateful for any suggestions. Where should I look for further clues?
On 09/27/2011 11:47 PM, Nick Urbanik wrote:
I have forced a selinux relabel, and a forcefsck.
Why? Nothing you report later mentions any SELinux denials or warnings and there's no mention of problems mounting partitions. I don't know what went wrong, but I do know that trying things at random isn't going to help and only wastes valuable time.
Dear Joe,
On 28/09/11 09:05 -0700, Joe Zeff wrote:
On 09/27/2011 11:47 PM, Nick Urbanik wrote:
I have forced a selinux relabel, and a forcefsck.
Why? Nothing you report later mentions any SELinux denials or warnings and there's no mention of problems mounting partitions. I don't know what went wrong, but I do know that trying things at random isn't going to help and only wastes valuable time.
Thanks for your insights into the good use of my valuable time.
On 10/02/2011 08:58 PM, Nick Urbanik wrote:
Thanks for your insights into the good use of my valuable time.
I spent about 8 years or so doing telephone tech support. One thing I learned is that trying things at random isn't as effective as taking a moment to look at the symptoms, deciding what could be causing it and, almost as important, what couldn't cause it. In your case, there were no SELinux alerts and no boot messages about corrupt file systems, so I saw no reason for you to disable SELinux or use fsck, which is why I suggested that you'd be better off looking someplace else. If you don't want me to make any further suggestions, just say so and I'll be glad to have Thunderbird delete your messages unread to make sure I don't.
A word to the wise: a sarcastic response will be read by me as a request to ignore you in the future.
On Sun, 2011-10-02 at 23:14 -0700, Joe Zeff wrote:
I spent about 8 years or so doing telephone tech support. One thing I learned is that trying things at random isn't as effective as taking a moment to look at the symptoms, deciding what could be causing it and, almost as important, what couldn't cause it.
And years of participating in forums, online, has shown me that rarely can you help someone solve a problem without asking them questions.
Usually, the person has not provided enough information, or any information. Usually, what they need to do to fix the problem is dependent on their actual problem. The stereotypical Windows, "reinstall and reboot without thinking," approaches rarely ever fixes any problem other than a random disk crash that hit a vital area. Likewise, with the "turn off your firewall or other protective software," dumb solution.
Most faults have a real cause, and *it* needs seeing to.
On Wed, 2011-09-28 at 16:47 +1000, Nick Urbanik wrote:
Dear Folks,
After a power surge and a spontaneous reboot, my 4-core F15 x86_64 machine has been unhappy. I have forced a selinux relabel, and a forcefsck.
Symptoms:
- The graphical boot did not complete: it started Nagios, but cupsd failed to start, and gdm did not start up.
- I cannot log in on the text consoles; I can type the username, but no prompt appears for the password.
- Attempts to use sudo just hang.
- I can log in by ssh as either myself or as root.
- named is using "97.5%" CPU, though when I turn on querylogging, there are very few lookups.
- Attempts to start cups result in complaints about dbus:
# systemctl start cups.service Failed to get D-Bus connection: Failed to authenticate in time.
Wonder if running into a systemd problem that I had due to dbus issue day or 2 ago?
Might upgrade to this and see if that helps fix it.
http://koji.fedoraproject.org/koji/buildinfo?buildID=265279
Dear Mike,
On 28/09/11 11:42 -0500, Mike Chambers wrote:
On Wed, 2011-09-28 at 16:47 +1000, Nick Urbanik wrote:
Dear Folks,
After a power surge and a spontaneous reboot, my 4-core F15 x86_64 machine has been unhappy. I have forced a selinux relabel, and a forcefsck.
(The fsck was necessary since the power surge had damaged the file system.)
Wonder if running into a systemd problem that I had due to dbus issue day or 2 ago?
Might upgrade to this and see if that helps fix it.
I just returned to work today, and my machine was locked up, keyboard showing no response to caps lock, num lock. I rebooted and found that its behaviour was similar to that of my machine at home. I rebooted into run level 1, and installed these systemd packages, which unfortunately did not solve the issue for me.
I used the same work around I used at home, detailed in this post to systemd-devel: http://lists.freedesktop.org/archives/systemd-devel/2011-September/003537.ht...