* Warren Togami wtogami@redhat.com [2005-05-01 19:43]:
- x86_64 Eclipse trouble
========================= libgcj-4.0.0-2 eclipse-ecj-3.1.0_fc-0.M6.12
x86_64 eclipse on 2GHz Athlon64 3200+ takes about 25 seconds to start. /tmp on the x86_64 contains hundreds of zero size files like "workbench.jar.sox3zske.so" [warren@fedora64 tmp]$ ls -l *jar* |wc -l 1156
Hmm. I'm not sure what those files are. I'll find an x86_64 box tomorrow and look into it.
Has anybody on fedora-devel-java seen this before?
Andrew
On Sun, 1 May 2005, Andrew Overholt wrote:
- Warren Togami wtogami@redhat.com [2005-05-01 19:43]:
- x86_64 Eclipse trouble
========================= libgcj-4.0.0-2 eclipse-ecj-3.1.0_fc-0.M6.12
x86_64 eclipse on 2GHz Athlon64 3200+ takes about 25 seconds to start. /tmp on the x86_64 contains hundreds of zero size files like "workbench.jar.sox3zske.so" [warren@fedora64 tmp]$ ls -l *jar* |wc -l 1156
Hmm. I'm not sure what those files are. I'll find an x86_64 box tomorrow and look into it.
Has anybody on fedora-devel-java seen this before?
Yes.
Linux www 2.4.21-4.ELsmp #4 SMP Mon Feb 2 03:36:19 MST 2004 x86_64 x86_64 x86_64 GNU/Linux
# ls /tmp/*.so|wc -l 1093
Only ping worked after trying it as root. The system was hung. Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=152386
rpm -qa eclipse* glibc* |sort eclipse-bugzilla-0.1.0_fc-13 eclipse-cdt-3.0.0_fc-0.M6.5 eclipse-changelog-2.0.1_fc-21 eclipse-ecj-3.1.0_fc-0.M6.12 eclipse-jdt-3.1.0_fc-0.M6.12 eclipse-jdt-devel-3.1.0_fc-0.M6.12 eclipse-pde-3.1.0_fc-0.M5.13 eclipse-pde-3.1.0_fc-0.M6.12 eclipse-pde-devel-3.1.0_fc-0.M6.12 eclipse-platform-3.1.0_fc-0.M6.12 eclipse-platform-devel-3.1.0_fc-0.M6.12 eclipse-pydev-0.9.3_fc-7 glibc-2.3.4-18 glibc-2.3.4-18 glibc-common-2.3.4-18 glibc-devel-2.3.4-18 glibc-devel-2.3.4-18 glibc-headers-2.3.4-18 glibc-kernheaders-2.4-9.1.90 glibc-profile-2.3.4-18 glibc-utils-2.3.4-18
Thats more of a zoo than I thought it was :)
On Wed, May 04, 2005 at 06:05:52AM -0400, Trent Jarvi wrote: <snip>
Only ping worked after trying it as root. The system was hung. Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
Yeah, you're right, we haven't been giving it much attention but mostly because it's not a normal use case. We are trying to fix the bugs that will affect the most number of users first.
Cheers, Ben
* Trent Jarvi tjarvi@qbang.org [2005-05-04 08:34]:
[...] Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
How does everyone feel about checking for root in /usr/bin/eclipse? If the user runs it from the menu system, what will happen in that case? Is there an easy way to pop up a message?
Andrew
* Andrew Overholt overholt@redhat.com [2005-05-13 11:19]:
- Trent Jarvi tjarvi@qbang.org [2005-05-04 08:34]:
[...] Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
How does everyone feel about checking for root in /usr/bin/eclipse? If
After further thought, I think this is a silly bandaid solution.
Andrew
Andrew Overholt (overholt@redhat.com) said:
- Trent Jarvi tjarvi@qbang.org [2005-05-04 08:34]:
[...] Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
How does everyone feel about checking for root in /usr/bin/eclipse? If the user runs it from the menu system, what will happen in that case? Is there an easy way to pop up a message?
I would think it's just simpler to figure out why it's hanging the system and fix that... did I miss that part of the discussion?
Bill
* Bill Nottingham notting@redhat.com [2005-05-13 11:35]:
Andrew Overholt (overholt@redhat.com) said:
- Trent Jarvi tjarvi@qbang.org [2005-05-04 08:34]:
[...] Maybe a check should be put in for root as an application hanging the system is this a critical bug reproducable with current devel i686 kernels and is not currently recieving attention.
How does everyone feel about checking for root in /usr/bin/eclipse? If the user runs it from the menu system, what will happen in that case? Is there an easy way to pop up a message?
I would think it's just simpler to figure out why it's hanging the system and fix that... did I miss that part of the discussion?
Andrew Haley thinks it's a kernel bug since the only way gij could hang the system is by forking very fast (which we don't think is happening here).
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=152386
Andrew
java-devel@lists.fedoraproject.org
-
Andrew Overholt -
Benjamin Konrath -
Bill Nottingham -
Trent Jarvi