When running cmdb_update_ambiguity on over 50,000 CIs; the domsrvr crashes. You will see similar issues as such in the logs:
12/05 11:08:09.92 sd-thojo22-lab domsrvr 2380 FATAL init.c 363 Operator ::new cannot allocate 4224 bytes
12/05 11:08:19.17 sd-thojo22-lab sqlagt:select3 3796 SIGNIFICANT sqlclass.c 1046 The following statement took 2601 milliseconds:
Clause (SELECT ci_superseded.last_update_date , ci_superseded.last_update_user , ci_superseded.not_ambiguous , ci_superseded.superseded_by FROM ci_superseded WHERE ci_superseded.id = ?)
12/05 11:08:19.27 sd-thojo22-lab kt_daemon 4000 ERROR main.c 203 Domsrvr connection lost
12/05 11:08:19.28 sd-thojo22-lab animator_nxd 2308 ERROR animator_nxd.c 635 The domsrvr has died! Reconnect will be attempted in 60 seconds
This is product limitation. It is advised to split the update in 25,000 or less. Ideally run it after loading CIs(grloader or GUI) if only less than 50,000 CIs are loaded or updated. The problem is that the domsrvr memory reaches the Operating system limit and that is why it dies. This is a known problem. Furthermore, it is known problem even with grloader loading over 50,000 CIs.
For official document regarding grloader limitation please review: RI27896 - https://support.ca.com/irj/portal/anonymous/redirArticles?reqPage=search&searchID=RI27896.