On Tue, 2005-02-01 at 18:38 -0500, Sam Varshavchik wrote:
> Mike Benoit writes:
>
> > In my limited experience, it appears to be the same. Very strange. It
> > happens pretty regularly now too.
> >
> > FAILED:
> > ----------------
> > 3720 stat64(".", {st_dev=makedev(8, 3), st_ino=1439300,
> > st_mode=S_IFDIR|S_ISUID|S_ISGID|0775, st_nlink=7, st_u
> > id=102, st_gid=8, st_blksize=4096, st_blocks=8, st_size=4096,
> > st_atime=2005/02/01-10:52:14, st_mtime=2005/02/01
> > -03:55:24, st_ctime=2005/02/01-03:55:24}) = 0
> > 3720 chdir("ipso_sm") = -1 EACCES (Permission denied)
> > 3720 fstat64(1, {st_dev=makedev(0, 4), st_ino=148341, st_mode=S_IFSOCK|
> > 0777, st_nlink=1, st_uid=0, st_gid=0, s
> > t_blksize=1024, st_blocks=0, st_size=0, st_atime=0, st_mtime=0,
> > st_ctime=0}) = 0
> >
> >
> > WORKED:
> > ----------------
> >
> > 3762 stat64(".", {st_dev=makedev(8, 3), st_ino=1439300,
> > st_mode=S_IFDIR|S_ISUID|S_ISGID|0775, st_nlink=7, st_u
> > id=102, st_gid=8, st_blksize=4096, st_blocks=8, st_size=4096,
> > st_atime=2005/02/01-10:52:14, st_mtime=2005/02/01 -03:55:24,
> > st_ctime=2005/02/01-03:55:24}) = 0
> > 3762 chdir("ipso_sm") = 0
> > 3762 stat64("loginexec", 0xbffff74c) = -1 ENOENT (No such file or
> > directory)
> > 3762 write(2, "INFO: LOGIN, user=ipso_sm, ip=[::ffff:207.81.249.35]\n",
> > 53) = 53
> > 3762 open("tmp", O_RDONLY|O_NONBLOCK|O_LARGEFILE|O_DIRECTORY) = 3
> > 3762 fstat64(3, {st_dev=makedev(8, 3), st_ino=1439310, st_mode=S_IFDIR|
> > S_ISGID|0770, st_nlink=2, st_uid=102, s
> > t_gid=8, st_blksize=4096, st_blocks=8, st_size=4096,
> > st_atime=2005/02/01-11:05:40, st_mtime=2005/02/01-11:05:40
> > , st_ctime=2005/02/01-11:05:40}) = 0
>
> There are only two possibilities:
>
> A) Filesystem corruption. Fsck everything.
> Its a brand new server, just installed a couple days ago. This issue doesn't only happen with one account either, it happens at random with every account that I have tried now. (about 20) My only guess at this point is an issue with my hardware Raid1 setup. Perhaps the reads are being distributed from drive to drive, and one drive has an issue? Seems pretty far fetched that I wouldn't be seeing similar issues with MySQL, or PGSQL, or LDAP, or Apache (which are all running also), if this was the case though. I wonder if bad ram could be causing it... Again though, weird that I'm only seeing issues with Courier. > B) A transient problem with your authentication module. You mentioned you > use MySQL. Check the logs for any errors. > I don't see any errors in the logs. But isn't the chdir() call well past the MySQL auth portion of the code? If MySQL was returning bad results, or failing in some manner, Courier should never get to this point correct? -- Mike Benoit <[EMAIL PROTECTED]>
signature.asc
Description: This is a digitally signed message part
