Bug #58409 (critical) network-backend.c:163: time went backwards

Description:
Hi, dear

I use crontab
0 * * * *    ( /usr/sbin/ntpdate time.nist.gov > /dev/null 2>&1 )
to adjust the system time of my rhel as5.

mysql-proxy.log tails with a lots of line as below as soon as minute 0 comes( see crontab above):

...
2010-11-23 13:08:37: (critical) network-backend.c:163: time went backwards (-2 usec)!
# many lines like above line
...

At the same time, the apps via the proxy work normally.

As you know, system time must be adjusted as accurate as possible.

So what I need to know from you is whether these critical log-lines imply me that the proxy has been crushed somehow and became more unstable?

Or could you please publish a log-error-reference?

Thanks.

How to repeat:
# run this then take a look at mysql-proxy.log
# rhel as5
/usr/sbin/ntpdate time.nist.gov

Thank you for the report.

Verified as described mostly feature request "publish log-error-reference" as message itself seems to be correct in this situation.

Hi, Sveta

My poor English told me that you must answer me as simply as possible. o(︶︿︶)o 

Q1: Whether or not those critical messages are warning me an unstable status of the proxy? 

if they arn't, see Q2.
else see Q3.

Q2: Those critical messages should not be classified in "critial" if they are not so "critial", aren'they? They make me difficult to find messages I really want to take a look. (my log-level = message)

Q3: What shall I do with this?

Thanks.

Hi, 

A new test showing:

Sometimes when these messages append the log, a lot of TCP connections from php-fgi to myproxy established and holding there( normally they disappear quickly ), so the users can feel a slow-down when they visit the pages.