[opensuse] File Read Time Reversion on XFS?
Hi, Recently I noticed what is to me is a very strange phenomenon: Read times on files are reverting to earlier values. The way I know (and noticed) this is that I read all the files in a particular directory at a specific time in order to monitor when another user later read those files. What I saw was that this directory listing: % ls -ltu total 176 -rwxrwxr-x 1 jhalcomb users 1033 2010-01-15 10:20 analyze* -rw-rw-r-- 1 jhalcomb users 3346 2010-01-15 10:20 defs-analyzed -rw-rw-r-- 1 jhalcomb users 3176 2010-01-15 10:20 defs-declared -rw-rw-r-- 1 jhalcomb users 212 2010-01-15 10:20 defs-pseudo -rw-rw-r-- 1 jhalcomb users 382 2010-01-15 10:20 defs-undeclared -rw-rw-r-- 1 jhalcomb users 641 2010-01-15 10:20 NIST-psl_outer_core-axm.sig -r--r--r-- 1 jhalcomb users 20549 2010-01-15 10:20 NIST-psl_outer_core.clif -rw-rw-r-- 1 jhalcomb users 4526 2010-01-15 10:20 NIST-psl_outer_core.defs -rw-rw-r-- 1 jhalcomb users 582 2010-01-15 10:20 NIST-psl_outer_core-def.sig -rw-rw-r-- 1 jhalcomb users 14765 2010-01-15 10:20 NIST-psl_outer_core-frag-ad.in -rw-rw-r-- 1 jhalcomb users 14542 2010-01-15 10:20 NIST-psl_outer_core-frag.in -rw-rw-r-- 1 jhalcomb users 14335 2010-01-15 10:20 NIST-psl_outer_core.in -rw-rw-r-- 1 jhalcomb users 710 2010-01-15 10:20 NIST-psl_outer_core.sig -rw-rw-r-- 1 jhalcomb users 9374 2010-01-15 10:20 NIST-psl_outer_core.toc -rw-rw-r-- 1 jhalcomb users 9002 2010-01-15 10:20 psl-1016-subset-618-lemmas.clif -rw-rw-r-- 1 jhalcomb users 1881 2010-01-15 10:20 psl28-theories -rwxrwxr-x 1 jhalcomb users 5947 2010-01-15 10:20 pslfrag* -rwxrwxr-x 1 jhalcomb users 8717 2010-01-15 10:20 pslprob* -rwxrwxr-x 1 jhalcomb users 5363 2010-01-15 10:20 pslsub* -rwxrwxr-x 1 jhalcomb users 161 2010-01-15 10:20 showfrags* -rwxrwxr-x 1 jhalcomb users 163 2010-01-15 10:20 showfrags-ad* ... later became this, without me doing anything: % ls -ltu total 176 -rwxrwxr-x 1 jhalcomb users 8717 2010-01-15 09:33 pslprob* -rwxrwxr-x 1 jhalcomb users 5947 2010-01-15 08:08 pslfrag* -rwxrwxr-x 1 jhalcomb users 5363 2010-01-15 08:08 pslsub* -rwxrwxr-x 1 jhalcomb users 1033 2010-01-13 13:22 analyze* -rw-rw-r-- 1 jhalcomb users 1881 2010-01-13 13:22 psl28-theories -rw-rw-r-- 1 jhalcomb users 4526 2010-01-13 13:15 NIST-psl_outer_core.defs -rw-rw-r-- 1 jhalcomb users 582 2010-01-13 13:15 NIST-psl_outer_core-def.sig -rw-rw-r-- 1 jhalcomb users 641 2010-01-13 13:15 NIST-psl_outer_core-axm.sig -rw-rw-r-- 1 jhalcomb users 710 2010-01-13 13:15 NIST-psl_outer_core.sig -rw-rw-r-- 1 jhalcomb users 9374 2010-01-13 13:15 NIST-psl_outer_core.toc -rw-rw-r-- 1 jhalcomb users 14765 2010-01-13 13:15 NIST-psl_outer_core-frag-ad.in -rw-rw-r-- 1 jhalcomb users 14542 2010-01-13 13:15 NIST-psl_outer_core-frag.in -rw-rw-r-- 1 jhalcomb users 14335 2010-01-13 13:15 NIST-psl_outer_core.in -rwxrwxr-x 1 jhalcomb users 163 2010-01-13 12:26 showfrags-ad* -rwxrwxr-x 1 jhalcomb users 161 2010-01-13 10:54 showfrags* -r--r--r-- 1 jhalcomb users 20549 2010-01-12 20:10 NIST-psl_outer_core.clif -rw-rw-r-- 1 jhalcomb users 212 2010-01-09 14:08 defs-pseudo -rw-rw-r-- 1 jhalcomb users 382 2010-01-09 14:07 defs-undeclared -rw-rw-r-- 1 jhalcomb users 3176 2010-01-09 14:06 defs-declared -rw-rw-r-- 1 jhalcomb users 3346 2010-01-09 14:03 defs-analyzed -rw-rw-r-- 1 jhalcomb users 9002 2009-04-21 10:27 psl-1016-subset-618-lemmas.clif Is this a known behavior of XFS? Or does it indicate some other process manipulating the read times on these files? (By the way, that other user I mention, the only other user on this system, is not sufficiently Unix- / Linux-savvy to deliberately manipulate file read times and has not used the system at all over the time period between the first and second directory listings shown above.) Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Friday, 2010-01-15 at 14:20 -0800, Randall R Schulz wrote:
Recently I noticed what is to me is a very strange phenomenon: Read times on files are reverting to earlier values. The way I know (and noticed) this is that I read all the files in a particular directory at a specific time in order to monitor when another user later read those files. What I saw was that this directory listing:
Mount options, perhaps?
% ls -ltu total 176
-t sort by modification time -u with -lt: sort by, and show, access time with -l: show access time and sort by name otherwise: sort by access time ...
Is this a known behavior of XFS? Or does it indicate some other process manipulating the read times on these files?
Beagle or whatever is the new kde equivalent. But the times would be newer, not older. - -- Cheers, Carlos E. R. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) iEYEARECAAYFAktQ9K0ACgkQtTMYHG2NR9V+bgCgk5vA+UuWSpJgg1fVo0qTPfQ6 8y0AnjiUziGA9XUjnAuxnY2hCrVDehOt =9Y+s -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
On Friday January 15 2010, Carlos E. R. wrote:
On Friday, 2010-01-15 at 14:20 -0800, Randall R Schulz wrote:
Recently I noticed what is to me is a very strange phenomenon: Read times on files are reverting to earlier values. The way I know (and noticed) this is that I read all the files in a particular directory at a specific time in order to monitor when another user later read those files. What I saw was that this directory listing:
Mount options, perhaps?
% ls -ltu total 176
-t sort by modification time -u with -lt: sort by, and show, access time with -l: show access time and sort by name otherwise: sort by access time
...
Is this a known behavior of XFS? Or does it indicate some other process manipulating the read times on these files?
Beagle or whatever is the new kde equivalent. But the times would be newer, not older.
I do use Google Desktop (but not Beagle), but as you say, that would cause the read time to change to a later time. Plus Google Desktop does not index the directory in question. My current suspicion falls on CrashPlan, my backup software.
-- Cheers, Carlos E. R.
Randall Schulz -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org
participants (2)
-
Carlos E. R.
-
Randall R Schulz