Summary: | evolution-data-server not killed on logout, preventing reboot. | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | stefan.ihringer |
Component: | [OLD] GNOME | Assignee: | Gentoo Linux Gnome Desktop Team <gnome> |
Status: | RESOLVED TEST-REQUEST | ||
Severity: | normal | CC: | kernel |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
URL: | http://forums.gentoo.org/viewtopic.php?t=250817&start=0&postdays=0&postorder=asc&highlight= | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 72798 |
Description
stefan.ihringer
2004-11-14 08:16:50 UTC
this isn't critical... what version of data-server ? figure out why it hangs.. evolution-data-server-1.0.2, see bug report above :-) I'd like to figure out why it hangs. But how'd I do that? I'm gonna downgrade to evo 2.0.1 and eds 1.0.1. Downgrading to evolution-2.0.1 and evolution-data-server-1.0.1 didn't help. Behaviour is the same. What else can I do to find out why it hangs? and you say it does work alright with other kernels ? No, that turned out to be a misperception. The update to evolution 2.0 must have happened at the same time I had upgraded my kernel. I figured out a work-around: In /etc/conf.d/local.stop I added the line "killall evolution-data-server-1.0". you could attach gdb to the e-d-s process and see if that comes up with anything on why it hangs ? Be sure to recompile libs with CFLAGS="-g" thought & ditch -fomit-frame-pointer completely, it screws up debugging for virtually no gain. I guess I have experienced the same problem. linux-2.6.5-gentoo-r1: System halts OK. linux-2.6.9-gentoo-r3: System hangs on readonly remount. In list of hanging processes is also spamd called as user. I guess it is kernel or sysvinit killall5 problem. foser: I don't have any experience debugging stuff. I played around with gdb a bit, attaching it to a process. It stopped the process and I figured out how to resume it. I don't know how I would have to use gdb to tell me what's wrong when eds is being killed. Sorry... I experience the same problem, but do not have any solution. can the kernel team have a look here if they got any indication what could've changed kernel-wise to this effect. Notice the clear upgrade indication in comment #7 . I myself am a development-sources-2.6.9 user and do not experience the problem. Could any of the reporters here try the stock 2.6.9 kernel to see if it's a gentoo specific patch ? Quite possibly the inotify stuff we added ... apparently thats now used for GNOME stuff, it's not in mainline kernel yet but we apply it as a patch it's not inotify if it uses 0.15 or higher. That is the one patch i do apply myself to the stock kernel. Current releases include 0.14 Don't suppose any of the 3 who report this as a bug can try patching their kernels with inotify 0.14? and does anything show up in your logs? I meant 0.15 of course, maybe I should just sleep. Have a pointer to the inotify 0.15 patch? Care to attach it here? Actually, 0.16 is out at http://www.kernel.org/pub/linux/kernel/people/rml/inotify/v2.6/0.16/inotify-0.16-rml-2.6.10-rc2-2.patch Do you think this one will work instead? As it is against 2.6.10-rc2, the odds of it working properly in 2.6.9 is pretty slim. That is why we include 0.14, it's the last one availble for 2.6.9 0.15 originally was patched also vs. 2.6.9 (gamin archives are down, can't give a link) those .16 patches seem to be rml's own and afaik there hasn't been an announced 0.16 release. I do know that one of rml .16's seemed to give me problems with NFS, but i haven't investigated yet. But yeah, i'm still waiting for someone who has this problem try these patches and assert if the inotify patches are the problem to begin with. I agree, closing with test-request. Hang in sync command in mount_readonly() in halt.sh has disappeared after upgrade from gentoo-dev-sources-2.6.9-r6 to gentoo-dev-sources-2.6.10-r1. I have not yet tested, whether there are any remaining tasks in this point. |