{ "date": "2017-02-01T13:06:33", "slug": "running-slob-on-exadata-express-cloud-service", "link": "https://www.dbi-services.com/blog/running-slob-on-exadata-express-cloud-service/", "title": { "rendered": "Running SLOB on Exadata Express Cloud Service" }, "content": { "rendered": "

By Franck Pachot

\n

.
\nThe Exadata Express Cloud Service is a managed PDBaaS: Oracle is the system admin and the CDB database administrator, you are the PDB administrator. You connect with a local user, PDB_ADMIN, which is no SYSDBA privilege but has a PDB_DBA which has nearly all DBA rights, but with some features disabled by lockdown profile.
\nI have no worry about the performance on this service: it is an Exadata X5 half rack bare metal (my guess), the CDB running as RAC One Node on 2 compute nodes, accessing the 7 storage cells. Smart Scan is disabled and given the small size of the PDB, and the low usage of the CDB, I/O is fast (most single block latency bwtween 128 and 256us coming from cell flash cache).
\nBut I like to run SLOB on new platforms and I wanted to see if I can run it here, without SYSDBA role, and connecting to a CDB.
\nYou can find SLOB at https://kevinclosson.net/slob/
\n

\n

TNS_ADMIN

\n

I’ll connect remotely because SLOB will be running on my computer (I have no access to the host for the PDBaaS managed service). In order to connect remotely, through a SQL*Net over SSL, you have to download the client credentials kit
\n\"CaptureEXCS017\"

\n

It is a zip file that contains the encryption wallet and the sqlnet.ora and tnsnames.ora to use. I unzip it in a directory (/var/tmp/client_credentials in this example) and I’ll use it by setting the TNS_ADMIN environment to this directory.
\nThus I change the wallet location in the sqlnet.ora:
\n$ cat sqlnet.ora

\n
WALLET_LOCATION = (SOURCE = (METHOD = file)\n                   (METHOD_DATA = (DIRECTORY=\"$TNS_ADMIN\")))\n
\n

and I use the provided tnsnames.ora which defines the ‘dbaccess’ service:

\n
$ cat tnsnames.ora\ndbaccess = (description=\n          (address=(protocol=tcps)(port=1522)(host=dbaccess.us2.oraclecloudapps.com))\n          (connect_data=(service_name=eptdojjm1ag.usdc2.oraclecloud.com))\n          (security=(ssl_server_cert_dn=\"CN=dbaccess.us2.oraclecloudapps.com,O=Oracle Corporation,L=Redwood Shores,ST=California,C=US\"))\n       )
\n

Changes in slob.conf

\n

The default slob.conf connects with a bequeath connection. Here I need to connect through the network service described above. I uncommented:

\n
\n#ADMIN_SQLNET_SERVICE=slob\n#SQLNET_SERVICE_BASE=slob\n#SYSDBA_PASSWD=change_on_install\n
\n

and replaced it with:

\n
\nADMIN_SQLNET_SERVICE=dbaccess\nSQLNET_SERVICE_BASE=dbaccess\nSYSDBA_PASSWD=xxxxxxxxx\n
\n

Where xxxxxxxxx is my PDB_ADMIN password that I setup in:
\n\"CaptureEXCS024\"

\n

Of course we need an Oracle Client. Download instant client if you don’t have one.

\n

Changes in setup.sh

\n

SLOB creates the users and grant them RESOURCE,DBA
\nIn Exadata Express Cloud Service, I don’t have the DBA role but the CREATE SESSION, and the PDB_ROLE which is sufficient. In addition to that I need to give quota to the IOPS tablespace because PDB_ADMIN do not have unlimited tablespace. Finally here is the create user part after my modifications:

\n
\nmsg NOTIFY \"Creating user: $user \"\n \nsqlplus \"$constring\" <<EOF\nWHENEVER SQLERROR EXIT 2;\nSET TERMOUT ON\nSET ECHO ON\nPROMPT User grants for $user\nWHENEVER SQLERROR EXIT 2;\nGRANT CREATE SESSION TO $user IDENTIFIED BY $user;\nGRANT PDB_DBA TO $user;\nALTER USER $user DEFAULT TABLESPACE $tablespace ;\nALTER USER $user QUOTA UNLIMITED ON $tablespace ;\nEXIT;\nEOF\n
\n

The setup.sh starts by dropping the SLOB schemas up to MAX_SLOB_SCHEMAS just in case they are there. The default is 4096 but that’s too much for me because of the latency to the server in US (it is planned to have Exadata Express Cloud Service in Europe in the future), so I reduced it to 8 (as I have only 1 CPU in this service, there’s no need for more users):

\n
\nexport MAX_SLOB_SCHEMAS=8\n
\n

And last modification in setup.sh was to disable the Multitenant checking because PDBaaS is by definition on a Multitenant database:

\n
\nif ( ! check_mto \"$ADMIN_CONNECT_STRING\")\nthen\n        msg FATAL \"\"\n        msg FATAL \"This version of SLOB does not support Oracle Multitenant Option\"\n        msg FATAL\n        #exit 1\nfi\n
\n

I’ve not seen any problem. Oracle designed multitenant so that everything you used to do on a database is possible to do on a PDB.

\n

Changes in runit.sh

\n

First, in order to connect as PDB_ADMIN instead of SYS as SYSDBA, you have to change the admin_connect_string:

\n
\n        export admin_connect_string=\"pdb_admin/${SYSDBA_PASSWD}@${ADMIN_SQLNET_SERVICE}\"\n
\n

Second, SLOB calls UTL_FILE which is not allowed in Exadata Express Cloud Service. Becase I’ve seen in the code that this is bypassed when the ‘nt’ argument is 1 so I’ve set it to this value in runit.sh:

\n
\n        arg9=1 #was: $nt\n
\n

Finally, because we can’t create manual AWR snapshots in Exadata Express Cloud Service, I replaced the EXEC DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT by the creation of a small view to get the basic statistics:

\n
\ncreate or replace view FPAV as\nselect 'STAT' class,name,value from V$SYSSTAT\nunion all\nselect 'WAIT' class,event,total_waits from V$SYSTEM_EVENT\nunion all\nselect 'TIME' class,stat_name,value from V$SYS_TIME_MODEL\nunion all\nselect 'V$OS' class,stat_name,value from V$OSSTAT\nunion all\nselect 'MICRO' class,event||' - '||wait_time_format,wait_count from V$EVENT_HISTOGRAM_MICRO\n;\ncreate table FPAT as select sysdate time,FPAV.* from FPAV where 1=0;\ninsert into FPAT select sysdate time,FPAV.* from FPAV;\ncommit;\n
\n

and I replaced the part that runs the AWR report with a simple query:

\n
\n($admin_conn <<EOF\nset echo on heading on underline on linesize 120 trimspool on pagesize 1000;\nhost cat slob.conf > awr.txt\nspool awr.txt append\ncommit;\ncolumn per_sec format 999G999G999G999G999\nselect to_char(btim,'hh24:mi:ss') btime,to_char(time,'hh24:mi:ss')etime,seconds,class,name,round(delta/seconds) per_second , decode(class||' '||name\n ,'STAT session logical reads','LIO/s'\n ,'STAT physical read total IO requests','PIO/s'\n ,'TIME DB CPU','CPU us'\n ,'TIME DB time','DB time us'\n ,'STAT physical read total bytes','Read B/s'\n ,'STAT physical write total bytes','Write B/s'\n ,'V$OS BUSY_TIME','OS busy cs'\n ,'WAIT resmgr:cpu quantum','resmgr'\n) main from (\n select\n FPAT.*\n ,max(time)over(order by time rows between unbounded preceding and unbounded following) last\n ,value-lag(value)over(partition by class,name order by time) delta\n ,(time-lag(time)over(partition by class,name order by time))*24*60*60 seconds\n ,lag(time)over(partition by class,name order by time) btim\n from FPAT\n) where time=last and delta>0 order by main,per_second desc;\nspool off\nhost cp awr.txt awr$(date +%Y%m%d%H%M%S).txt\nexit\nEOF\n) &\n
\n

Statspack is not an alternative here because it requires some views and grants from SYS which you cannot do on Exadata Express Cloud Service. I really don’t understand why we cannot use AWR locally because in 12.2 you can have AWR snapshots at PDB level. But any use of dbms_workload_repository is disabled by lockdown profile.

\n

Result

\n

So this is the kind of output I get on running 2 SLOB sessions during 10 minutes:

\n
\nBTIME    ETIME       SECONDS CLAS NAME                                                             PER_SECOND MAIN\n-------- -------- ---------- ---- ---------------------------------------------------------------- ---------- ----------\n11:16:09 11:26:15        606 TIME DB CPU                                                              1968151 CPU us\n11:16:09 11:26:15        606 TIME DB time                                                             1981479 DB time us\n11:16:09 11:26:15        606 STAT session logical reads                                               1228557 LIO/s\n11:16:09 11:26:15        606 STAT physical read total IO requests                                           0 PIO/s\n11:16:09 11:26:15        606 STAT physical read total bytes                                               189 Read B/s\n11:16:09 11:26:15        606 STAT physical write total bytes                                              189 Write B/s\n11:16:09 11:26:15        606 WAIT resmgr:cpu quantum                                                        0 resmgr\n11:16:09 11:26:15        606 STAT logical read bytes from cache                                    1.0064E+10\n11:16:09 11:26:15        606 TIME sql execute elapsed time                                            1980376\n11:16:09 11:26:15        606 V$OS IDLE_TIME                                                              6787\n11:16:09 11:26:15        606 V$OS BUSY_TIME                                                               391\n11:16:09 11:26:15        606 V$OS USER_TIME                                                               311\n
\n

1.96 CPU seconds per second and 1.98 DB time. Those are my 2 cached SLOB sessions.
\nAbout 1.2 million LIO per seconds. This is what I expect from those test on Exadata X5 (PCT_UPDATE=0 and WORK_UNIT=64).
\nResource manager ready to kick in as my CPU_COUNT is set to 2 in my PDB (set by lockdown profile, you cannot change it)
\nThe overall system hosting the CDB has been mostly idle: 3.91 / (67.87 + 3.91) is 5.44% busy. I’m responsible for 2 sessions over 72 threads (36 cores): 2/72=2.78% CPU usage

\n

The Exadata Express Cloud Service allocates 1 core (2 threads) but because the system is nearly idle, I didn’t use 2 threads of the same core. Let’s compare with a 1 session only run:

\n
\nBTIME    ETIME       SECONDS CLAS NAME                                                             PER_SECOND MAIN\n-------- -------- ---------- ---- ---------------------------------------------------------------- ---------- ----------\n12:15:49 12:25:55        606 TIME DB CPU                                                               989008 CPU us\n12:15:49 12:25:55        606 TIME DB time                                                              991198 DB time us\n12:15:49 12:25:55        606 STAT session logical reads                                                668368 LIO/s\n12:15:49 12:25:55        606 STAT physical read total IO requests                                           0 PIO/s\n12:15:49 12:25:55        606 STAT physical read total bytes                                               203 Read B/s\n12:15:49 12:25:55        606 STAT physical write total bytes                                              189 Write B/s\n12:15:49 12:25:55        606 STAT logical read bytes from cache                                    5475272359\n12:15:49 12:25:55        606 TIME sql execute elapsed time                                             990290\n12:15:49 12:25:55        606 STAT consistent gets                                                      668368\n
\n

Comparing logical reads, 1228557/668368=1.8 so probably I was not running the two sessions on the same core.

\n

So what?

\n

SLOB (https://kevinclosson.net/slob/) doesn’t need to be installed on the database server because it is mostly PL/SQL, so no roundtrips. And this is a way to run it when you have only access to a PDB without SYSDBA privileges.

\n", "protected": false } }