Deadlock condition in Directory Server Retro Changelog plug-in causes IPA to become unresponsive

Issue

A customer executed a host deletion operation (ipa host-del) which subsequently caused the IPA server to become unresponsive. An inspection of the Directory Server error log to determine the root cause of the problem revealed a deadlock condition in the Retro changelog plug-in (retrocl-plugin) :

[25/Apr/2016:12:55:39 +0000] - Retry count exceeded in delete
[25/Apr/2016:12:55:39 +0000] DSRetroclPlugin - delete_changerecord: could not delete change record 2107045 (rc: 51)
[25/Apr/2016:12:57:03 +0000] - Operation error fetching Null DN (3424e306-0ae511e6-a817c0c2-8e7cec37), error -30993.
[25/Apr/2016:12:57:03 +0000] - dn2entry_ext: Failed to get id for changenumber=2137810,cn=changelog from entryrdn index (-30993)
[25/Apr/2016:12:57:03 +0000] - Operation error fetching changenumber=2137810,cn=changelog (null), error -30993.
[25/Apr/2016:12:57:03 +0000] DSRetroclPlugin - replog: an error occured while adding change number 2137810, dn = changenumber=2137810,cn=changelog: Operations error.
[25/Apr/2016:12:57:03 +0000] retrocl-plugin - retrocl_postob: operation failure [1]
[25/Apr/2016:12:57:05 +0000] - ldbm_back_seq deadlock retry BAD 1601, err=0 BDB0062 Successful return: 0
[25/Apr/2016:12:57:24 +0000] - slapd shutting down - signaling operation threads - op stack size 11 max work q size 9 max work q stack size 9
[25/Apr/2016:12:57:24 +0000] - slapd shutting down - waiting for 29 threads to terminate