Bug 1452397

Description of problem:

SSSD fails to retrieve all the users and groups intermittently when there are large number of users and groups when enumeration is set to true.

In my case, customer has 12000 users and 1300 groups.
  

Version-Release number of selected component (if applicable):

sssd-1.14.0-43.el7_3.14


How reproducible:

1. Integrate RHEL 7.3 machine through SSSD to AD which has users more than 10,000 users and groups more than 1000. 

2. Enable enumeration by adding enumerate = true in sssd.conf. 

3. Set-up a cronjob which runs every hour a script to check how many 
passwd and group entries are available with getent and redirect the result to a file. 

Cronjob executes the below script every hour : 

pwe=$(getent passwd |wc -l); gre=$(getent group | wc -l ); if 
[ $pwe -lt 10000 -o $gre -lt 1000 ]; then myd="$(date 
+\%Y-\%m-\%dT\%H:\%M)"; echo "$myd FAIL pwe=$pwe and gre=$gre" | tee -a 
/tmp/enumlog; fi

Actual results:

File has some entries when getent returns users less than 10,000 or groups less than 1000. 

cat /tmp/enumlog

2017-04-24T20:12 FAIL pwe=78 and gre=103
2017-04-25T03:12 FAIL pwe=27 and gre=60
2017-04-25T10:12 FAIL pwe=69 and gre=133
2017-04-25T14:12 FAIL pwe=11286 and gre=421
2017-04-25T17:12 FAIL pwe=54 and gre=115
2017-04-26T00:12 FAIL pwe=30 and gre=66
2017-04-26T07:12 FAIL pwe=30 and gre=63
2017-04-26T14:12 FAIL pwe=61 and gre=128
2017-04-26T18:12 FAIL pwe=11689 and gre=342
2017-04-26T21:12 FAIL pwe=30 and gre=67
2017-04-27T04:12 FAIL pwe=27 and gre=60

Expected results:

getent should return correct number of users and groups entries every time. 


Additional Information : 

I will collect SSSD debug logs and will upload shortly.