Red Hat Bugzilla – Bug 1000113
race condition during startup
Last modified: 2013-08-23 11:11:52 EDT
Created attachment 789283 [details]
Description of problem:
Crash during startup because some blivet calls that use g-i-introspection started before X did
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Started a virt-install using Fedora-19-x86_64-DVD.iso
2. Crash before the welcome screen.
In addition to the attached logs, the following was on tty1:
anaconda 19.30.13-1 for Fedora 19 started
[xcb] Unknown sequence number while processing queue
[xcb] Most likely this is a multi-threaded client and XInitThreads has not been called
[xcb] Aborting, sorry about that.
python: xcb_io.c:274: poll_for_event: Assertion `!xcb_xlib_threads_sequenceLost' failed.
Created attachment 789284 [details]
Created attachment 789286 [details]
Created attachment 789290 [details]
Created attachment 789291 [details]
Created attachment 789292 [details]
Created attachment 789294 [details]
Created attachment 789295 [details]
Created attachment 789296 [details]
Created attachment 789297 [details]
The exception string is always the best choice for an initial bug summary:
AttributeError: 'NoneType' object has no attribute '__name__'
"race condition during startup" is your interpretation and not at all informative.
It is also useful to paste error messages from the logs into a comment, so they can be found by searching BZ:
Error message from attached anaconda.log:
17:34:51,190 ERR anaconda: Problems running the window manager: [Errno 0] SIGCHLD caught when trying to start the X server.
The log you have attached as "traceback" appears to be an anaconda-tb* file. Please use the exact name of the file, so that it can be recognized without opening it.
(In reply to Steve Tyler from comment #12)
> The log you have attached as "traceback" appears to be an anaconda-tb* file.
> Please use the exact name of the file, so that it can be recognized without
> opening it.
This is the name you should use: "anaconda-tb-p_QLAy":
BTW, you can change it by clicking Details, Edit Details for the attachment.
(In reply to Steve Tyler from comment #10)
> The exception string is always the best choice for an initial bug summary:
> AttributeError: 'NoneType' object has no attribute '__name__'
> "race condition during startup" is your interpretation and not at all
You can change the bug summary by clicking the "edit" link next to the bug summary at the top of the bug page.
(In reply to David Shea from comment #0)
> How reproducible:
> Obnoxiously rarely
What is this supposed to mean?
Have you seen it more than once?
Steve, please stop telling people who work on anaconda how to file and handle bugs in anaconda. Your behavior is unhelpful and frankly quite rude.
Steve, traceback is just as descriptive as anaconda-tb-* as far as what's actually in the file. I really don't care what the output of mktemp was.
Yes, race condition is my interpretation.
It would be helpful if you would consolidate these comments into a single comment next time. Brevity is the soul of wit, but that doesn't count when multiplied by six.
Chris, get out of your little I_Am_Anaconda world. Other people have to read bug reports to reproduce and analyze the bugs.
Dave, are your colleagues giving you any mentoring? Those faults in your bug report are so elementary, I was surprised to see an RH employee making them.
FYI, mktemp ensures uniqueness of names. When you are doing triage, maybe you have noticed that, sometimes, more than one anaconda-tb* file is attached. If everyone followed your one-name-fits-all approach, there could be several attachments, all named "traceback".
Also, you should state just the facts in your initial report and put your interpretation in followup comments. That way, the initial report is not biased by a possibly mistaken interpretation.
Here is some take-home reading:
Bug Writing Guidelines
This bug is being closed. All further comments to it will be ignored.
Communication in this bug report is in violation of the Fedora Code of Conduct: