[arch-general] Can't create /tmp/.X11-unix and /tmp/.ICE-unix while booting
Just before I do something stupid, I thought I would ask about these apparent errors. Mon May 30 11:49:39 2011: ^[[1;34;40m:: ^[[1;37;40mRemoving Leftover Files^[[1;0m ^[[s^[[151G ^[[1;34;40m[^[[0;36;40mBUSY^[[1;34;40m]^[[1;0m /bin/mkdir: cannot create directory `/tmp/.X11-unix': File exists Mon May 30 11:49:40 2011: /bin/mkdir: cannot create directory `/tmp/.ICE-unix': File exists My boot log goes back to 12 May 2011 so that's the earliest I can verify this data. I've search flyspray for any bug reports about this and didn't find any. Should these two files be deleted on shutdown or should they not be created on boot? If they should be deleted, where is the error? Myra -- Life's fun when your sick and psychotic!
On Mon, May 30, 2011 at 8:26 PM, Myra Nelson <myra.nelson@hughes.net> wrote:
Just before I do something stupid, I thought I would ask about these apparent errors.
Mon May 30 11:49:39 2011: ^[[1;34;40m:: ^[[1;37;40mRemoving Leftover Files^[[1;0m ^[[s^[[151G ^[[1;34;40m[^[[0;36;40mBUSY^[[1;34;40m]^[[1;0m /bin/mkdir: cannot create directory `/tmp/.X11-unix': File exists Mon May 30 11:49:40 2011: /bin/mkdir: cannot create directory `/tmp/.ICE-unix': File exists
My boot log goes back to 12 May 2011 so that's the earliest I can verify this data. I've search flyspray for any bug reports about this and didn't find any. Should these two files be deleted on shutdown or should they not be created on boot? If they should be deleted, where is the error?
The bug report you are looking for is this one: <https://bugs.archlinux.org/task/24279>. It was closed as the fix is already in git (but not yet released): <http://projects.archlinux.org/initscripts.git/commit/?id=cc1b8a39309fef45d092140fa130d0e82c696a44>. Summary: the change was unintentionally introduced by me when changing some nearby code. It only affects people who have /tmp mounted on a non-tmpfs partition (people who just leave /tmp on the rootfs are not affected). The fix will be included in the next release. Since I already have your attention, I'd like to suggest you do the following: Turn the partiton that is currently holding your /tmp into swapspace Mount /tmp as tmpfs setting size= the size of your old /tmp partition. In addition to solving the bug in question, it will give you better performance, save power (fewer writes to disk) and increase the amount of available swap, without increasing the RAM usage. For a discussion of why this is almost always the right thing to do (unless you happen to have files larger than 256GB on your /tmp), have a look at the above bug report. Cheers, Tom
On Mon, May 30, 2011 at 13:41, Tom Gundersen <teg@jklm.no> wrote:
On Mon, May 30, 2011 at 8:26 PM, Myra Nelson <myra.nelson@hughes.net> wrote:
Just before I do something stupid, I thought I would ask about these apparent errors.
Mon May 30 11:49:39 2011: ^[[1;34;40m:: ^[[1;37;40mRemoving Leftover Files^[[1;0m ^[[s^[[151G ^[[1;34;40m[^[[0;36;40mBUSY^[[1;34;40m]^[[1;0m /bin/mkdir: cannot create directory `/tmp/.X11-unix': File exists Mon May 30 11:49:40 2011: /bin/mkdir: cannot create directory `/tmp/.ICE-unix': File exists
My boot log goes back to 12 May 2011 so that's the earliest I can verify this data. I've search flyspray for any bug reports about this and didn't find any. Should these two files be deleted on shutdown or should they not be created on boot? If they should be deleted, where is the error?
The bug report you are looking for is this one: <https://bugs.archlinux.org/task/24279>. It was closed as the fix is already in git (but not yet released): <http://projects.archlinux.org/initscripts.git/commit/?id=cc1b8a39309fef45d092140fa130d0e82c696a44>.
Summary: the change was unintentionally introduced by me when changing some nearby code. It only affects people who have /tmp mounted on a non-tmpfs partition (people who just leave /tmp on the rootfs are not affected). The fix will be included in the next release.
Since I already have your attention, I'd like to suggest you do the following:
Turn the partiton that is currently holding your /tmp into swapspace Mount /tmp as tmpfs setting size= the size of your old /tmp partition.
In addition to solving the bug in question, it will give you better performance, save power (fewer writes to disk) and increase the amount of available swap, without increasing the RAM usage.
For a discussion of why this is almost always the right thing to do (unless you happen to have files larger than 256GB on your /tmp), have a look at the above bug report.
Cheers,
Tom
Tom Thanks, I'll take care of the changes to /tmp. My /tmp is on a separate partition and I seriously hope there are no file on my /tmp partition even remotely close to 256GB. The only reason I have a /tmp as large as I do is for extra space when compiling something and a separte /tmp partition is to prevent any possible race conditions from outside influences. I'll always accept any help I can get. Old age has dented my memory and it makes things harder to keep up with. Myra -- Life's fun when your sick and psychotic!
Hi Myra, On Mon, May 30, 2011 at 9:02 PM, Myra Nelson <myra.nelson@hughes.net> wrote:
Thanks, I'll take care of the changes to /tmp. My /tmp is on a separate partition and I seriously hope there are no file on my /tmp partition even remotely close to 256GB. The only reason I have a /tmp as large as I do is for extra space when compiling something and a separte /tmp partition is to prevent any possible race conditions from outside influences.
Using tmpfs should work well for you then (shout if it doesn't I'd be interested to know!). You could also apply the patch that I linked to, if you prefer to keep your current setup. Or you could wait for a new initscript release, which shouldn't be too far off :-) Cheers, Tom
On Mon, May 30, 2011 at 14:17, Tom Gundersen <teg@jklm.no> wrote:
Hi Myra,
On Mon, May 30, 2011 at 9:02 PM, Myra Nelson <myra.nelson@hughes.net> wrote:
Thanks, I'll take care of the changes to /tmp. My /tmp is on a separate partition and I seriously hope there are no file on my /tmp partition even remotely close to 256GB. The only reason I have a /tmp as large as I do is for extra space when compiling something and a separte /tmp partition is to prevent any possible race conditions from outside influences.
Using tmpfs should work well for you then (shout if it doesn't I'd be interested to know!).
You could also apply the patch that I linked to, if you prefer to keep your current setup. Or you could wait for a new initscript release, which shouldn't be too far off :-)
Cheers,
Tom
Tom Seems to work fine. I'll think I'l wait on the /var/run and /var/run/lock until the changes hit the repos. rather than change all the rc.d scripts. Again, thanks for your help. -- Life's fun when your sick and psychotic!
On Mon, May 30, 2011 at 9:50 PM, Myra Nelson <myra.nelson@hughes.net> wrote:
Seems to work fine. I'll think I'l wait on the /var/run and /var/run/lock until the changes hit the repos. rather than change all the rc.d scripts.
Again, thanks for your help.
Happy to help :-) Cheers, Tom
participants (2)
-
Myra Nelson
-
Tom Gundersen