View previous topic :: View next topic |
Author |
Message |
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3444 Location: Gainesville, Florida
|
Posted: Mon Aug 07, 2006 12:34 am Post subject: Strange superblock boot message (mount time in future) |
|
|
I've suddenly been getting random boot messages like this on all my ext3 partitions:
superblock last mount time in future FIXED
I read about a possible cause on a slackware forum:
Quote: | e2fsprogs version 1.39 is messed up on boot as far as the time
goes.If you set your system time utc everything is ok,but if you set it
local time you get messages on boot |
Anybody seeing this, and have a fix? I refuse to use utc time- prefer local.
Is it something to be concerned about?
I'm using 2.6.17-beyond2.2, no4, 2.6.18-rc3-viper1, and 2.6.18-rc2-mm1 kernels. Doesn't seem to matter what kernel I'm using- I still get these messages, but not every boot. Nothing else seems amiss- my system is running perfectly. _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.9.1 USE=experimental python3_11
Last edited by wrc1944 on Wed Aug 16, 2006 10:16 pm; edited 1 time in total |
|
Back to top |
|
|
suicidal_orange_II Apprentice
Joined: 04 Sep 2004 Posts: 299
|
Posted: Mon Aug 07, 2006 2:16 pm Post subject: |
|
|
I'm guessing you live somewhere where the timezone is gmt +x, in which case if you reboot in less than x hours you will see this message.
It is not going to do any damage to your data so I wouldnt worry about it (although a fix would be nice )
Suicidal_Orange |
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3444 Location: Gainesville, Florida
|
Posted: Mon Aug 07, 2006 7:07 pm Post subject: |
|
|
Nope- I'm in Florida, gmt -5. This is only happening on one of my three boxes (so far).
I thought maybe one of these kernels was the culprit, but I compiled then on all boxes, and that's not it, so it's a local thing on one box. All the systems are identical, i.e. config files are all the same- no clue there. Guess I'll check the bios time settings and see if that has somehow gone out of whack. These are pretty new boards, so the battery should be perfectly fine. _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.9.1 USE=experimental python3_11 |
|
Back to top |
|
|
Section_8 l33t
Joined: 22 May 2004 Posts: 627
|
Posted: Wed Aug 16, 2006 12:18 pm Post subject: |
|
|
I noticed that message rebooting lastnight. I am in CST (gmt-6). I also have my hardware clock set to local time, for windows dual boot of course. |
|
Back to top |
|
|
Ivar_Y n00b
Joined: 29 Jun 2004 Posts: 25
|
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3444 Location: Gainesville, Florida
|
Posted: Thu Aug 17, 2006 12:36 pm Post subject: |
|
|
Hmmm- I read through bug 142850, and the last comment #11 by Ivar seems like it's probably the solution.
Quote: | The /etc/init.d/checkfs startup script generates the "FIXED" error messages
and, as Gentoo Linux is currently configured, it runs before /etc/init.d/clock.
A solution is to run /etc/init.d/clock before /etc/init.d/checkfs. The order
that these scripts run in is controlled by the "CRITICAL_SERVICES" list in
/sbin/rc. (According to the Gentoo Handbook, initscripts in /etc/init.d are
supposed to run in alphabetical order as modified by "before" and "after"
dependencies but, on my system, this doesn't seem to work.) Hence,
CRITICAL_SERVICES="checkroot modules checkfs localmount clock bootmisc" needs
to be changed to CRITICAL_SERVICES="checkroot modules clock checkfs localmount
bootmisc". The current startup scripts indicate that "clock" needs
"localmount" and that "localmount" needs "checkfs." While the latter
dependency seems reasonable, the former does not. "Clock" corrects the system
time and this does not obviously depend on file system checks or on mounting
the files in /etc/fstab. (Root is already mounted.) |
So if I understand correctly, those of us having this problem should change line 167 in sbin/rc to:
CRITICAL_SERVICES="checkroot modules clock checkfs localmount bootmisc"
Can anyone confirm this? I'm still having trouble understanding why it seems to happen randomly, and not all the time, and more often depending on what kernel you are booted to. All time related options are the same in all my kernels on all boxes (at least I'm pretty certain they are).
I also can't understand that if Ivar's solution is correct, why has Gentoo been configured differently for years, and nobody has noticed or mention this before? Or is the issue suddenly completely new because of something else that has recently changed (i.e., e2fsprogs-1.39? _________________ Main box- AsRock x370 Gaming K4
Ryzen 7 3700x, 3.6GHz, 16GB GSkill Flare DDR4 3200mhz
Samsung SATA 1000GB, Radeon HD R7 350 2GB DDR5
OpenRC Gentoo ~amd64 plasma, glibc-2.36-r7, gcc-13.2.1_p20230304
kernel-6.9.1 USE=experimental python3_11
Last edited by wrc1944 on Thu Aug 17, 2006 3:22 pm; edited 1 time in total |
|
Back to top |
|
|
Section_8 l33t
Joined: 22 May 2004 Posts: 627
|
Posted: Thu Aug 17, 2006 2:44 pm Post subject: |
|
|
My understanding from reading that bug is that e2fsprogs 1.39 does a last mount time sanity check that earlier versions didn't do. So the startup local/UTC time confusion was always there, it just never produced an error message.
I'll try that fix when I get home tonight. |
|
Back to top |
|
|
Ivar_Y n00b
Joined: 29 Jun 2004 Posts: 25
|
Posted: Thu Aug 17, 2006 3:25 pm Post subject: |
|
|
To add to Apprentice's post:
From the "e2fsprogs-1.39" RELEASE-NOTES:
Quote: | E2fsck will detect if the superblock's last mount field or last write
field is in the future, and offer to fix if so. (Addresses Debian Bug
#327580) These problems will be fixed automatically in preen mode
since Debian's boot sequence bogusly doesn't set the time correctly
until potentially very late in the bootup process, and this can cause
false positives which will cause users' systems to fail to boot.
(Addresses Debian Bugs #343662 and #343645) |
The check is not present in "e2fsprogs-1.38."
A possible explanation of differences among boxes is differences in the settings of "/etc/localtime" in the boxes. Do the results of a "date"command show the same time zone on each of the boxes? I have two computers. One has "/etc/localtime" linked to "/usr/share/zoneinfo/US/Eastern." "/etc/localtime" did not exist on the second. The "date" command showed an "EDT" zone for the first, a "UTC" zone for the second. When I added the "/etc/localtime" link, the "time in the future" problem appeared and both showed the "EDT" zone.
Regardless, the "time in the future" error message never appears for my "home" partition on my primary computer. It is not obvious to me why. The "debugfs" command shows that the last mount and last write times are four hours earlier than they should be.
Ivar |
|
Back to top |
|
|
yabbadabbadont Advocate
Joined: 14 Mar 2003 Posts: 4791 Location: 2 exits past crazy
|
Posted: Thu Aug 17, 2006 10:22 pm Post subject: |
|
|
Instead of modifying /sbin/rc, I created /etc/runlevels/boot/.critical and put the new service order there. It is used if it exists. I then had to move the clock service before checkroot as the root check would always show the error (while the other filesystems wouldn't). As was stated earlier, clock doesn't rely on any other services and the root filesystem is already mounted. Also, I removed the depend() function from the clock service. The change to /etc/init.d/clock is the only thing I will have to be sure gets preserved whenever baselayout is updated. _________________
Bones McCracker wrote: | On the other hand, regex is popular with the ladies. |
|
|
Back to top |
|
|
Ivar_Y n00b
Joined: 29 Jun 2004 Posts: 25
|
Posted: Fri Aug 18, 2006 4:49 am Post subject: |
|
|
Putting "clock" at the front of the CRITICAL_SERVICES list is a good idea. Putting it after checkroot and before checkfs means that "last mount time" and "last write time" are still set to wrong values for the root file system. This is probably unimportant but starting the list off with clock assures that these values are accurate.
I do not need to remove the depend() function in clock.
I suspect that keeping the CRITICAL_SERVICES list in /sbin/rc is a better idea than keeping it as /etc/runlevels/boot/.critical. I assume that users will have to maintain and update the list in the latter location and this is likely to confuse many of them. /sbin/rc is a part of baselayout, which is a Gentoo package maintained by the Gentoo Developers. It seems safer to let them keep things straight.
Ivar |
|
Back to top |
|
|
yabbadabbadont Advocate
Joined: 14 Mar 2003 Posts: 4791 Location: 2 exits past crazy
|
Posted: Fri Aug 18, 2006 7:56 pm Post subject: |
|
|
I removed the depend function since it only contained "need localmount". (which it doesn't)
Regarding maintaining the .critical file after baselayout updates. I guess it's a "six of one and half a dozen of the other" situation. Either you will have to maintain the .critical file if the list of services changes, or you will have to keep editing the /sbin/rc file after each baselayout update that doesn't have the clock service in the correct place. Take your pick. (of course, once baselayout has a correct /sbin/rc, neither option will be needed) _________________
Bones McCracker wrote: | On the other hand, regex is popular with the ladies. |
|
|
Back to top |
|
|
csab Apprentice
Joined: 15 Apr 2005 Posts: 152 Location: Atlanta, GA
|
Posted: Fri Sep 22, 2006 1:37 am Post subject: |
|
|
Developers screwing us... with the new baselayout (1.12.5) the following lines were added to /sbin/rc:
Code: |
# Start checkroot and modules before we load volumes
export START_CRITICAL="yes"
for x in checkroot modules ; do
[[ " ${CRITICAL_SERVICES} " != *" ${x} "* ]] && continue
start_critical_service "${x}"
done
|
This kills our workaround. I understand that the addition fixes a bug with RAID, but isn't that true that a lot more people use local time than RAID?
I'll hack /sbin/rc for now and I'll watch the development of bug #142850. |
|
Back to top |
|
|
yabbadabbadont Advocate
Joined: 14 Mar 2003 Posts: 4791 Location: 2 exits past crazy
|
Posted: Fri Sep 22, 2006 2:07 am Post subject: |
|
|
Yeah. I ended up doing the same thing. _________________
Bones McCracker wrote: | On the other hand, regex is popular with the ladies. |
|
|
Back to top |
|
|
UberLord Retired Dev
Joined: 18 Sep 2003 Posts: 6835 Location: Blighty
|
Posted: Fri Sep 22, 2006 6:42 am Post subject: |
|
|
csab wrote: | Developers screwing us... with the new baselayout (1.12.5) the following lines were added to /sbin/rc |
We're not screwing you.
However to please you, baselayout-1.13_alpha1 should be hitting portage sometime next week as it feature integrated vserver and more importantly fbsd support. How does this affect you? Well, we've also done away with the concept of a "critical service" as we can now run ALL services using their dependencies.
So you'll just be able to modify the deps in the init scripts. _________________ Use dhcpcd for all your automated network configuration needs
Use dhcpcd-ui (GTK+/Qt) as your System Tray Network tool |
|
Back to top |
|
|
csab Apprentice
Joined: 15 Apr 2005 Posts: 152 Location: Atlanta, GA
|
Posted: Fri Sep 22, 2006 2:55 pm Post subject: |
|
|
UberLord wrote: | csab wrote: | Developers screwing us... with the new baselayout (1.12.5) the following lines were added to /sbin/rc |
We're not screwing you. |
C'mon, I didn't mean it! Sorry if I offended you, guys. Next time I will write a smiley.
And thanks for the information! |
|
Back to top |
|
|
csab Apprentice
Joined: 15 Apr 2005 Posts: 152 Location: Atlanta, GA
|
Posted: Sun Mar 16, 2008 10:49 pm Post subject: |
|
|
UberLord wrote: | csab wrote: | Developers screwing us... with the new baselayout (1.12.5) the following lines were added to /sbin/rc |
We're not screwing you.
However to please you, baselayout-1.13_alpha1 should be hitting portage sometime next week as it feature integrated vserver and more importantly fbsd support. How does this affect you? Well, we've also done away with the concept of a "critical service" as we can now run ALL services using their dependencies.
So you'll just be able to modify the deps in the init scripts. |
Isn't it funny. I came back to this thread and I saw this post. Dear developer promised baselayout 1.13 alpha for next week on 2006 Sept. 22. It's a YEAR AND A HALF later, and the stable baselayout is still 1.12. (And the next version, 2.0 is hard masked.) Needless to say that the bug is still there.
Something is pretty wrong around Gentoo. |
|
Back to top |
|
|
overkll Veteran
Joined: 21 Sep 2004 Posts: 1249 Location: Austin, Texas
|
Posted: Sun May 10, 2009 5:11 am Post subject: |
|
|
Too funny, thought I'd be the only one digging this up again. I had to dig this up out of my email to find this bug/thread again. Indeed it's still an issue. I use UTC on machines when I can, but my lappy needs "local". I couldn't remember what I did the last time. This fix works for me:
Code: | --- /sbin/rc.orig 2009-04-05 14:29:31.000000000 -0500
+++ /sbin/rc 2009-05-09 23:48:17.000000000 -0500
@@ -436,7 +436,7 @@
# Start checkroot and modules before we load volumes
export START_CRITICAL="yes"
- for x in checkroot modules ; do
+ for x in clock checkroot modules ; do
[[ " ${CRITICAL_SERVICES} " != *" ${x} "* ]] && continue
start_critical_service "${x}"
done
@@ -449,7 +449,7 @@
# We do not want to break compatibility, so we do not fully integrate
# these into /sbin/rc, but rather start them by hand ...
for x in ${CRITICAL_SERVICES} ; do
- [[ ${x} == "checkroot" || ${x} == "modules" ]] && continue
+ [[ ${x} == "clock" || ${x} == "checkroot" || ${x} == "modules" ]] && continue
start_critical_service "${x}"
done |
|
|
Back to top |
|
|
|