| Summary: | udevd always crashes after some cauldron updates | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Malo Deniélou <pmdenielou> |
| Component: | RPM Packages | Assignee: | Mageia Bug Squad <bugsquad> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | mageia, mageia, xrg |
| Version: | Cauldron | ||
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | udevd | CVE: | |
| Status comment: | |||
| Attachments: |
urpmi log of the installed packages. One of them is responsible.
dmseg output |
||
|
Description
Malo Deniélou
2012-06-30 02:00:15 CEST
Is this is a new thing noticed with the landing of new systemd+udev? Or has this been occurring for a while - your wording seems to suggest it's been happening for some time? It has been happening for a while, but it's the first time I thought about looking at the dmesg, which helped me link it to the systemd update. I'm on IRC to give more info if needed, although I don't know where to start ... Created attachment 2531 [details]
urpmi log of the installed packages. One of them is responsible.
It just happened again tonight, while updating a few packages (including systemd) and while no other program was running (meaning I wasn't doing anything on the machine except updating). I attach the list of packages that were updated this time.. I also attach my dmesg output which shows that udevd starts crashing at some point during the update. The crash is similar to the one mentioned above. It affects the network and interprocess communication: I observe that the network (firefox) is blocked and that e.g. the kde menu bar is unresponsive, until a new udevd process is launched (takes a dozen seconds) at which point all the actions are flushed (menu, network). It works for a couple of seconds, then it crashes again. Created attachment 2532 [details]
dmseg output
Happened again with a plymouth update today. According to https://bbs.archlinux.org/viewtopic.php?id=144270 It's linked to a firmware problem. But I don't think so, since after just a reboot, everything works fine. I will however try to get a new firmware, or just remove the wifi card altogether, and see if it helps for next update ... Ok. I can't signal any more crashes of udevd. It is because of the firmware. I installed ralink-firmware and it went away. It should not happen though. Any idea to prevent that? I have had the same trouble: In a new machine with *Mageia 2* (not cauldron), the b43 firmware was missing[1]. This caused the udevd to segfault[2]. I guess, by removing some firmware, this bug must be reproducible.. But there is another issue here: introduction of systemd causes this udevd loop to run more aggressively than before. This means, udevd keeps crashing and crashing, thus bringing the system down to its knees[3]. In my case, it would repeatedly reconfigure the SATA channel[4], raising the system load to >4.0 Can we also put a "respawn limit" or sth. similar to systemd for udevd? [1] as expected, since I had to get it from unofficial sites. [2] this bug, to be fixed. [3] because the detection loop of udevd slows down several resources. [4] I could see some sda-related messages all the time CC:
(none) =>
xrg do you still reproduce this bug ? CC:
(none) =>
nicolas.lecureuil I haven't seen that bug recently on cauldron or mageia 3. then let close it and please reopen if the issue reappear Status:
NEW =>
RESOLVED |