Hi,
I am not totally sure what caused this but I just
wanted to know if anything of this sort has been
seen in the ext3 world.
I did a fresh install of redhat 7.2 linux with
kernel-2.4.9(we have some patches applied to this).
I have root and boot filesystems as ext3 type. After
running some tests the system gets into a strange
state. I cannot move one file to another if that
destination file already exists. Here is some text
that can help you understand the problem:
.....
{lng}usersg_16: ls
{lng}usersg_17: touch aaa
{lng}usersg_18: ls -l
total 0
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 aaa
{lng}usersg_19: touch bbb
{lng}usersg_20: ls -l
total 0
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 aaa
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 bbb
{lng}usersg_21: mv aaa bbb
total 0
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 aaa
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 bbb
{lng}usersg_21: mv aaa ccc ------>(destination file doesn't exist)
{lng}usersg_22: ls -l
total 0
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 bbb
-rw-rw-r-- 1 usersg media 0 Sep 19 10:45 ccc
.....
The strace of the mv command shows that everything is done
correct and exits with code 0. A copy of file aaa to bbb
succeeds but a move fails.
I was just wondering if anything of this sort is in any
way related to some state of ext3 filesystem. Yet another
observation is that this problem is gone with a reboot.
Thanks in advance,
Girish
Girish Sheelvant wrote:> Hi, > > I am not totally sure what caused this but I just > wanted to know if anything of this sort has been > seen in the ext3 world. > > I did a fresh install of redhat 7.2 linux with > kernel-2.4.9(we have some patches applied to this). > > I have root and boot filesystems as ext3 type. After > running some tests the system gets into a strange > state. I cannot move one file to another if that > destination file already exists. Here is some text > that can help you understand the problem:[SNIP]> The strace of the mv command shows that everything is done > correct and exits with code 0. A copy of file aaa to bbb > succeeds but a move fails. > > I was just wondering if anything of this sort is in any > way related to some state of ext3 filesystem. Yet another > observation is that this problem is gone with a reboot.Possibly it's the following bug: http://marc.theaimsgroup.com/?l=linux-kernel&m=101362147103329&w=2 also http://marc.theaimsgroup.com/?l=linux-kernel&m=101338613730184&w=2 regards, Juri -- If each of us have one object, and we exchange them, then each of us still has one object. If each of us have one idea, and we exchange them, then each of us now has two ideas.
Unfortunately, we did not check the kernel logs when the problem occurred. Since the patch that Juri pointed us to appears to have solved the problem we have not had an opportunity to go back and check if anything was in the kernel logs. We will attempt to recreate the error scenario as soon as we get a chance and will report our findings in this forum. ...Roger... -----Original Message----- From: Stephen C. Tweedie [mailto:sct@redhat.com] Sent: Monday, September 23, 2002 2:03 PM To: Juri Haberland Cc: Girish Sheelvant; ext3-users@redhat.com Subject: Re: problem moving files Hi, On Thu, Sep 19, 2002 at 05:36:47PM +0200, Juri Haberland wrote:> > The strace of the mv command shows that everything is done > > correct and exits with code 0. A copy of file aaa to bbb > > succeeds but a move fails. > > Possibly it's the following bug: > http://marc.theaimsgroup.com/?l=linux-kernel&m=101362147103329&w=2If that's the problem, we'd expect to see something in the kernel logs indicating an IO failure leading to the bad inode. Is there anything like that showing up in this case? Cheers, Stephen _______________________________________________ Ext3-users mailing list Ext3-users@redhat.com https://listman.redhat.com/mailman/listinfo/ext3-users