Bug 128986 (IT_44347)

1) We need the output of 'satellite-debug' mailed to us (by output I
  mean /tmp/satellite-debug.tar.bz2, not stdout)
                                                                     
                                                            
2) We need sample emails, including full headers, of the duplicates
                                                                     
                                                            
3) We need /var/log/maillog from the satellite during a time that
   duplicate emails were sent

also, please send the output of this sql.  First, run sqlplus:

sqlplus rhnsat/rhnsat@rhnsat

then, once in sql plus, execute these three queries and send the
results to us:

SELECT COUNT(*) FROM RHNEMAILADDRESS;
SELECT COUNT(*) FROM WEB_CONTACT;
SELECT COUNT(DISTINCT EMAIL) FROM WEB_USER_PERSONAL_INFO;

thanks

SQL output:

SQL> SELECT COUNT(*) FROM RHNEMAILADDRESS;
 
  COUNT(*)
----------
       101
 
SQL> SELECT COUNT(*) FROM WEB_CONTACT;
 
  COUNT(*)
----------
       101
 
SQL> SELECT COUNT(DISTINCT EMAIL) FROM WEB_USER_PERSONAL_INFO;
 
COUNT(DISTINCTEMAIL)
--------------------
                 101

I'll upload the other requested files as private.

okay at this point we've eliminated the theories we have internally on
the issue and are trying a second attempt to reproduce as well as
analyzing the code in question for flaws.  we will let you know ASAP
when we have further information or need more data from your satellite.

Customer says:

I changed my satellite sync cron job to run a single command:

satellite-sync

I was using a separate command line satellite-sync -c channel  for
each channel.

This change appears to have eliminated the problems associated with
this ticket  including redundant e-mails, bogus errata scheduling,
syncs taking a very long time, taskomatic running for hours on end,
and channel packages not showing for some errata.

I made the cron change for the weekend.  The saturday and sunday syncs
errored out with the same error, but the monday one ran through
completely and all looks good.  I will monitor the sync for the next
several days.

Ping.  Do we have an ETA for a fix for the, erhm, spam :-) when
syncing individual channels?

Assigning to the appropriate person and linking to rhn360sat
s/cturner/misa

I believe Cliff mentioned something about auto-errata.
This may be fixed in 360 with the refactoring of satellite-sync, but
since we don't have a reproducer it's hard to tell. Moving to ON_DEV,
QA may have a better chance to reproduce this. Todd, please coordinate
with cperry if he has more details about this.

QA push. {ON_DEV,QA_READY} --> ON_QA

There are too many comments and I cannot figure out what is the real
problem/issue.  Could you please provide a test plan or ways to
reproduce the error to make sure it has been fixed? Thanks!

We never managed to reproduce this in our environments.

I haven't seen any duplicate erratas...  Closing this bug as we never
were able to reproduce it in our evn.