Created attachment 18438 [details] the modifications I attempted to make Overview: During the business processing of the kdc service, it seems that some resources sdb_entry_free have not been fully released in order to obtain the sdb entry constructed for the hdb entry. Steps to Reproduce: 1) Continuously perform kdc business service processing, such as kinit. 2) Keep the kdc service running for a long time. Actual Results: The server resource memory is continuously occupied until 100%. Expected Results: Server memory resources should not be 100% available without business requests. other： I found that the server memory keeps rising while using the built-in kdc service of samba. Based on this issue, I used the Valgrind tool, and the main information in the Valgrind report is as follows: ==16375== 56 bytes in 4 blocks are indirectly lost in loss record 33 of 420 ==16375== at 0x4C29F73: malloc (vg_replace_malloc.c:309) ==16375== by 0xDA9754D: krb5_data_alloc (data.c:101) ==16375== by 0xDA97615: krb5_data_copy (data.c:149) ==16375== by 0xC9CD117: smb_krb5_copy_data_contents (krb5_samba.c:936) ==16375== by 0xA87E4C3: samba_kdc_fill_user_keys (db-glue.c:514) ==16375== by 0xA87F03C: samba_kdc_message2entry_keys (db-glue.c:842) ==16375== by 0xA88035C: samba_kdc_message2entry (db-glue.c:1594) ==16375== by 0xA881ED0: samba_kdc_fetch_krbtgt (db-glue.c:2466) ==16375== by 0xA882EF3: samba_kdc_fetch (db-glue.c:2964) ==16375== by 0x20F41929: hdb_samba4_fetch_kvno (hdb-samba4.c:153) ==16375== by 0x2116D9D1: fetch_it (common.c:1661) ==16375== by 0x2116DCB1: hdb_fetch_kvno (common.c:1785) The above is a brief message. After my investigation, I have some doubts about these locations. Has there been a memory leak? 1.entry->max_life = malloc(sizeof(*entry->max_life)); [ source4/kdc/db-glue.c：1396] 2.entry->max_renew = malloc(sizeof(*entry->max_renew));[source4/kdc/db-glue.c:1416] 3.ret = sdb_entry_set_etypes(entry);[source4/kdc/db-glue.c:1638] 4.ret = sdb_entry_set_session_etypes(entry, add_aes256, add_aes128, add_rc4);[source4/kdc/db-glue.c:1650] 5.entry->keys.len = 1;[source4/kdc/db-glue.c:1687] I hope you can give me a reply when you have time. As a student, I am deeply puzzled. The attachment is the modification I made when I discovered the problem. Thank you very much, Best regards,