What’s new with Oracle database 19.9 versus 19.8
This blogpost takes a look at the technical differences between Oracle database 19 RU 8 (july 2020) and RU 9 (october 2020). This gives technical specialists an idea of the differences, and gives them the ability to assess if the RU impacts anything.
Functions
code symbol names unique in version 19.8 versus 19.9 (top 5) -- so removed in the 19.9 RU NAME RESOLVE ANNOTATION ARCHIVE_OBJECT COUNT(*) -------------------------------------------------- -------------------------------------------------- -------------------------------------------------------------------------------- ------------------------------------------------------------ ---------- qsod (qs)od query system ?? 3 ZSTD ZSTD Zstandard. fast lossless compression algorithm 4 qjsn qjsn query json 5 Java Java java classes 6 qcpi qcpi query compile parse interim 6 code symbol names unique in version 19.9 versus 19.8 (top 10) -- so added in the 19.9 RU NAME RESOLVE ANNOTATION ARCHIVE_OBJECT COUNT(*) -------------------------------------------------- -------------------------------------------------- -------------------------------------------------------------------------------- ------------------------------------------------------------ ---------- qsod (qs)od query system ?? libclntst19.a:generic19_qsodax.o,libgeneric19.a:qsodax.o 8 mp_s (mp_)s multiple precision cryptography ?? libjavavm19.a:mpi.o 11 qcpi qcpi query compile parse interim libclntst19.a:generic19_qcpi7.o,libgeneric19.a:qcpi7.o 11 ksac (ksa)c kernel service asynchronous messages ?? libserver19.a:ksacl.o 13 jznO (jzn)O json ?? libxml19.a:jznoct.o,libclntst19.a:xml19_jznoct.o 15 qjsn qjsn query json libserver19.a:qjsn.o 15 jznt (jzn)t json ?? libxml19.a:jzntrans.o,libclntst19.a:xml19_jzntrans.o 19 jzno (jzn)o json ?? libxml19.a:jznoct.o,libclntst19.a:xml19_jznoct.o 20 ec_G (ec_)G elleptic curve cryptography ?? libjavavm19.a:ecl_gf.o 21 s_mp s_mp cryptography multiple precision libjavavm19.a:mpi.o 49
This is a poor, yet one of the only ways to understand what is going on inside the oracle executable.
There aren’t that many thing removed, which is understandable.
I show the the top 11 of function groups that are added, because their numbers are significant (more than 10).
In general, what I see is most functions are related to cryptography, which is understandable and logical in today’s world.
Additionally, there’s a lot of movement with JSON related libraries, this is also understandable, and probably a lot of the functions are related to DBMS_SODA.
Obviously, there is more changed as we can see below from the parameters (which are switches for bug fixes), and other changes, but the function changes doe not indicate huge groupings of functions to be added or removed outside of the groups indicated above.
Parameters
parameters unique in version 19.8 versus 19.9 NAME -------------------------------------------------- _eleventh_spare_parameter _nineteenth_spare_parameter _one-hundred-and-seventy-fourth_spare_parameter _one-hundred-and-sixty-eighth_spare_parameter _thirteenth_spare_parameter _twentieth_spare_parameter _twenty-eighth_spare_parameter _twenty-fifth_spare_parameter _twenty-first_spare_parameter _twenty-fourth_spare_parameter _twenty-seventh_spare_parameter _twenty-sixth_spare_parameter _twenty-third_spare_parameter parameters unique in version 19.9 versus 19.8 NAME -------------------------------------------------- _bug28482048_fb_log_max_rereads _bug29504103_runtime_index_key_length_check _bug30159581_cputime_limit_parallel_fptr _bug30159581_runtime_limit_parallel_fptr _bug30165506_enable_unified_memory _bug30165506_unified_pga_gran_leak_timeout _bug30165506_unified_pga_list_count _bug30165506_unified_pga_max_pool_size _bug31747989_cputime_limit_parallel_fptr _bug31747989_runtime_limit_parallel_fptr _ipacl_feature_control_flags _ldap_use_all_direct_groups_only _pdb_transition_clean_bg_delay
Once again, this is quite consistent, ‘spare’ parameters being swapped for lots of ‘_bugnnnnnnn_description’ parameters, and a few other underscore parameters.
There are no new documented parameters.
Waitevents
waitevents unique in version 19.8 versus 19.9 NAME ---------------------------------------------------------------------------------------------------- second spare wait event waitevents unique in version 19.9 versus 19.8 NAME ---------------------------------------------------------------------------------------------------- unified memory allocation
A new wait event was added with PSU 19.9! A ‘spare’ wait event was changed to ‘unified memory allocation’. I do not have more information about this, and have not heard about ‘unified memory’ before.
Sysstat
sysstat statistics unique to 19.8 versus 19.9 NAME ---------------------------------------------------------------------------------------------------- cell RDMA reads cell pmem cache read hits spare statistic 1 spare statistic 2 sysstat statistics unique to 19.9 versus 19.8 NAME ---------------------------------------------------------------------------------------------------- cell RDMA reads cell RDMA reads unused cell pmem cache read hits cell pmem cache read hits unused
A few statistics have been added! Quite logically this is related to RDMA and pmem (persistent memory). It looks like existing statistics have been moved to a new statistics number, and the ‘old’ ones have gotten an addition saying ‘unused’. It’s understandable and known that oracle is putting efforts in the area’s of RDMA, where it seems the usage of RDMA is expanded all the time to increase performance of RAC cross instance communication, and pmem is new and currently officially implemented for Exadata. I do not know if the above statistic is the exadata related pmem cache, or if this is a local machine pmem implementation.
DBA/CDB views
dba tables unique to 19.8 versus 19.9 dba tables unique to 19.9 versus 19.8 NAME ---------------------------------------------------------------------------------------------------- DBA_SUBSCR_DUR_REGISTRATIONS cdb tables unique to 19.8 versus 19.9 cdb tables unique to 19.9 versus 19.8 NAME ---------------------------------------------------------------------------------------------------- CDB_SUBSCR_DUR_REGISTRATIONS
It seems there is a DBA and CDB scope view added.
Frits Hoogland Weblog 