Mailing List Archive

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug
report.

http://bugzilla.quagga.net/show_bug.cgi?id=332


erik@ripe.net changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |erik@ripe.net




------- Additional Comments From rimondin@dia.uniroma3.it 2007-01-11 09:34 -------
If I configure bgpd to produce dumps at predictable times (in my specific case,
dump bgp routes-mrt filename 8h), strange collection instants are observed in
file creation times. The output of ls -l inside my collection directory follows:

-rw-r----- 1 foo foo 0 2007-01-02 01:41 rib.20070102.0041
-rw-r----- 1 foo foo 24387242 2007-01-02 09:01 rib.20070102.0800
-rw-r----- 1 foo foo 24382856 2007-01-02 17:00 rib.20070102.1558
-rw-r----- 1 foo foo 24422446 2007-01-03 01:00 rib.20070102.2358
-rw-r----- 1 foo foo 24391556 2007-01-03 09:00 rib.20070103.0758
-rw-r----- 1 foo foo 24380481 2007-01-03 17:00 rib.20070103.1558
-rw-r----- 1 foo foo 24400158 2007-01-04 01:00 rib.20070103.2358
-rw-r----- 1 foo foo 24392996 2007-01-04 09:00 rib.20070104.0758
-rw-r----- 1 foo foo 24434670 2007-01-04 17:00 rib.20070104.1558
-rw-r----- 1 foo foo 24431762 2007-01-05 01:00 rib.20070104.2358
-rw-r----- 1 foo foo 24436674 2007-01-05 09:00 rib.20070105.0758
-rw-r----- 1 foo foo 8228547 2007-01-05 16:59 rib.20070105.1558
-rw-r----- 1 foo foo 8764251 2007-01-05 17:00 rib.20070105.1559
-rw-r----- 1 foo foo 24457144 2007-01-06 01:00 rib.20070105.2359
-rw-r----- 1 foo foo 24449854 2007-01-06 09:00 rib.20070106.0758
-rw-r----- 1 foo foo 24450730 2007-01-06 17:00 rib.20070106.1558

As you can see, by looking at the file names, all the RIBs but the first two
appear to be collected about 2 minutes before the expected scheduled time. Even
more strange, sometimes it happens that two dumps are produced within a very
limited time span (15:58 and 15:59), possibly resulting in incomplete file
contents (look at the file size). On the other hand, it is weird to notice that
the second dump (08:00 in the above list) took place at the correct time instant.
The 1 hour time offset you observe between the file name timestamp and the
actual file creation time is normal and due to quagga running inside a User-Mode
Linux sandbox within which the timezone is set to UTC (our Linux box is set to CET).

The clock drift pattern you see has been observed in several experiments I have
performed, even after changing the dump generation rate (12h instead of 8h).
Conversely, BGP updates have always been collected at the expected time
instants, even after changing the collection rate.

I experienced this issue with both Quagga v0.99.5 (Debian package) and Quagga
v0.99.6 (compiled from scratch). Everything seems to go fine with the latest
stable release (0.98.6).



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.
_______________________________________________
Quagga-bugs mailing list
Quagga-bugs@lists.quagga.net
http://lists.quagga.net/mailman/listinfo/quagga-bugs