We are using OpenLDAP as a password backend and have found every so often that
the slapd process runs away. This was traced down to repeated queries (around
1200 per second) for an invalid uid from an smbd process.
An example slapd log entry:
Jul 15 13:25:19 myhost slapd: conn=3 op=369 SRCH
More tracking has found that this uid (testuser) was listed in a memberUid
attribute of a samba ldap group.
This is of course a problem with invalid data in our ldap database (should only
have valid users listed), but I think the smbd should accept the first "nope, no
such user" response and get on with other things :)
We are okay now (we've done some tidying), but I'm hoping that someone will have
a look at the code so I can sleep easier in the future!
Should be fixed with
Could you please try again with samba3.0.5rc1 ?
*** This bug has been marked as a duplicate of 1478 ***