Bobby Durrett's DBA Blog

Subscribe to Bobby Durrett's DBA Blog feed
Oracle database performance
Updated: 1 month 2 weeks ago

Understanding query slowness after platform change

Thu, 2016-05-12 14:54

We are moving a production database from 10.2 Oracle on HP-UX 64 bit Itanium to 11.2 Oracle on Linux on 64 bit Intel x86. So, we are upgrading the database software from 10.2 to 11.2. We are also changing endianness from Itanium’s byte order to that of Intel’s x86-64 processors. Also, my tests have shown that the new processors are about twice as fast as the older Itanium CPUs.

Two SQL queries stand out as being a lot slower on the new system although other queries are fine. So, I tried to understand why these particular queries were slower. I will just talk about one query since we saw similar behavior for both. This query has sql_id = aktyyckj710a3.

First I looked at the way the query executed on both systems using a query like this:

select ss.sql_id,
where ss.sql_id = 'aktyyckj710a3'
and ss.snap_id=sn.snap_id
and executions_delta > 0
order by ss.snap_id,ss.sql_id;

It had a single plan on production and averaged a few seconds per execution:

PLAN_HASH_VALUE END_INTERVAL_TIME         EXECUTIONS_DELTA Elapsed Average ms CPU Average ms IO Average ms Cluster Average ms Application Average ms Concurrency Average ms Average buffer gets Average disk reads Average rows processed
--------------- ------------------------- ---------------- ------------------ -------------- ------------- ------------------ ---------------------- ---------------------- ------------------- ------------------ ----------------------
      918231698 11-MAY-16 PM              195         1364.80228     609.183405    831.563728                  0                      0                      0          35211.9487             1622.4             6974.40513
      918231698 11-MAY-16 PM              129         555.981481     144.348698    441.670271                  0                      0                      0          8682.84496         646.984496             1810.51938
      918231698 11-MAY-16 PM               39         91.5794872     39.6675128    54.4575897                  0                      0                      0          3055.17949          63.025641             669.153846
      918231698 12-MAY-16 AM               35         178.688971     28.0369429    159.676629                  0                      0                      0          1464.28571              190.8             311.485714
      918231698 12-MAY-16 AM              124         649.370258     194.895944    486.875758                  0                      0                      0           13447.871         652.806452             2930.23387
      918231698 12-MAY-16 AM              168         2174.35909     622.905935    1659.14223                  0                      0             .001303571          38313.1548         2403.28571             8894.42857
      918231698 12-MAY-16 AM              213         3712.60403     1100.01973    2781.68793                  0                      0             .000690141          63878.1362               3951             15026.2066
      918231698 12-MAY-16 PM              221         2374.74486      741.20133    1741.28251                  0                      0             .000045249          44243.8914         2804.66063               10294.81

On the new Linux system the query was taking 10 times as long to run as in the HP system.

PLAN_HASH_VALUE END_INTERVAL_TIME         EXECUTIONS_DELTA Elapsed Average ms CPU Average ms IO Average ms Cluster Average ms Application Average ms Concurrency Average ms Average buffer gets Average disk reads Average rows processed
--------------- ------------------------- ---------------- ------------------ -------------- ------------- ------------------ ---------------------- ---------------------- ------------------- ------------------ ----------------------
     2834425987 10-MAY-16 PM               41         39998.8871     1750.66015    38598.1108                  0                      0                      0          50694.1463         11518.0244             49379.4634
     2834425987 10-MAY-16 PM               33         44664.4329     1680.59361    43319.9765                  0                      0                      0          47090.4848         10999.1818             48132.4242
     2834425987 11-MAY-16 AM                8          169.75075      60.615125      111.1715                  0                      0                      0             417.375                 92                2763.25
     2834425987 11-MAY-16 PM               11         14730.9611     314.497455    14507.0803                  0                      0                      0          8456.63636         2175.63636             4914.90909
     2834425987 11-MAY-16 PM                2           1302.774       1301.794             0                  0                      0                      0               78040                  0                  49013
     2834425987 11-MAY-16 PM                1           1185.321       1187.813             0                  0                      0                      0               78040                  0                  49013
     2834425987 11-MAY-16 PM               14         69612.6197     2409.27829     67697.353                  0                      0                      0          45156.8571         11889.1429             45596.7143
     2834425987 11-MAY-16 PM               16         65485.9254     2232.40963    63739.7442                  0                      0                      0          38397.4375         12151.9375             52222.1875
     2834425987 12-MAY-16 AM               61         24361.6303     1445.50141    23088.6067                  0                      0                      0          47224.4426         5331.06557              47581.918
     2834425987 12-MAY-16 AM               86         38596.7262     1790.56574    37139.4262                  0                      0                      0          46023.0349         9762.01163             48870.0465

The query plans were not the same but they were similar. Also, the number of rows in our test cases were more than the average number of rows per run in production but it still didn’t account for all the differences.

We decided to use an outline hint and SQL Profile to force the HP system’s plan on the queries in the Linux system to see if the same plan would run faster.

It was a pain to run the query with bind variables that are dates for my test so I kind of cheated and replaced the bind variables with literals. First I extracted some example values for the variables from the original system:

select * from 
(select distinct
sb.sql_id='aktyyckj710a3' and
order by 

Then I got the plan of the query with the bind variables filled in with the literals from the original HP system. Here is how I got the plan without the SQL query itself:

truncate table plan_table;

explain plan into plan_table for 
-- problem query here with bind variables replaced

set markup html preformat on

select * from table(dbms_xplan.display('PLAN_TABLE',

This plan outputs an outline hint similar to this:

      NO_ACCESS(@"SEL$5DA710D3" "VW_NSO_1"@"SEL$5DA710D3")
      OPT_PARAM('query_rewrite_enabled' 'false')

Now, to force aktyyckj710a3 to run on the new system with the same plan as on the original system I had to run the query on the new system with the outline hint and get the plan hash value for the plan that the query uses.

explain plan into plan_table for 
Plan hash value: 1022624069

So, I compared the two plans and they were the same but the plan hash values were different. 1022624069 on Linux was the same as 918231698. I think that endianness differences caused the plan_hash_value differences for the same plan.

Then we forced the original HP system plan on to the real sql_id using coe_xfr_sql_profile.sql.

-- build script to load profile

@coe_xfr_sql_profile.sql aktyyckj710a3 1022624069

-- run generated script


Sadly, even after forcing the original system’s plan on the new system, the query still ran just as slow. But, at least we were able to remove the plan difference as the source of the problem.

We did notice a high I/O time on the Linux executions. Running AWR reports showed about a 5 millisecond single block read time on Linux and about 1 millisecond on HP. I also graphed this over time using my Python scripts:

Linux db file sequential read (single block read) graph:


HP-UX db file sequential read graph:


So, in general our source HP system was seeing sub millisecond single block reads but our new Linux system was seeing multiple millisecond reads. So, this lead us to look at differences in the storage system. It seems that the original system was on flash or solid state disk and the new one was not. So, we are going to move the new system to SSD and see how that affects the query performance.

Even though this led to a possible hardware issue I thought it was worth sharing the process I took to get there including eliminating differences in the query plan by matching the plan on the original platform.



Our Linux and storage teams moved the new Linux VM to solid state disk and resolved these issues. The query ran about 10 times faster than it did on the original system after moving Linux to SSD.

HP Version:

------------------------- ---------------- ------------------ PM                        245         5341.99923 PM                        250         1280.99632 PM                        341         3976.65855 PM                        125         2619.58894


------------------------- ---------------- ------------------
16-MAY-16 AM              162         191.314809
16-MAY-16 AM              342         746.313994
16-MAY-16 AM              258         461.641705
16-MAY-16 PM              280         478.601618

The single block read time is well under 1 millisecond now that 
the Linux database is on SSD.

END_INTERVAL_TIME          number of waits ave microseconds 
-------------------------- --------------- ---------------- 
15-MAY-16 PM           544681       515.978687
16-MAY-16 AM           828539       502.911935
16-MAY-16 AM           518322       1356.92377
16-MAY-16 AM            10698       637.953543
16-MAY-16 AM              193       628.170984
16-MAY-16 AM              112        1799.3125
16-MAY-16 AM             1680       318.792262
16-MAY-16 AM              140       688.914286
16-MAY-16 AM             4837       529.759768
16-MAY-16 AM            16082       591.632508
16-MAY-16 AM           280927       387.293204
16-MAY-16 AM           737846        519.94157
16-MAY-16 AM          1113762       428.772997
16-MAY-16 PM           562258       510.357372


Categories: DBA Blogs

Comparing Common Queries Between Test and Production

Thu, 2016-05-05 13:58

The developers complained that their test database was so much slower than production that they could not use it to really test whether their batch processes would run fast enough when migrated to production. They did not give me any particular queries to check. Instead they said that the system was generally too slow. So, I went through a process to find SQL statements that they had run in test and that normally run in production and compare their run times. I thought that I would document the process that I went through here.

First I found the top 100 queries by elapsed time on both the test and production databases using this query:

column FORCE_MATCHING_SIGNATURE format 99999999999999999999

sum(ELAPSED_TIME_DELTA) total_elapsed
order by total_elapsed desc)
where rownum < 101;

The output looked like this:


Then I found the signatures that were in common between the two lists.

insert into test_sigs values (944718698451269965);
insert into test_sigs values (4634961225655610267);
insert into test_sigs values (15939251529124125793);
insert into prod_sigs values (3898230136794347827);
insert into prod_sigs values (944718698451269965);
insert into prod_sigs values (11160330134321800286);
select * from test_sigs
select * from prod_sigs;

This led to 32 values of FORCE_MATCHING_SIGNATURE which represented queries that ran on both test and production, except for the possible difference in constants.

Next I looked at the overall performance of these 32 queries in test and production using this query:

create table common_sigs

insert into common_sigs values (575231776450247964);
insert into common_sigs values (944718698451269965);
insert into common_sigs values (1037345866341698119);

sum(executions_delta) total_executions,
from DBA_HIST_SQLSTAT ss,common_sigs cs

Here is part of the output:

TOTAL_EXECUTIONS Elapsed Average ms CPU Average ms IO Average ms
---------------- ------------------ -------------- -------------
         5595295         366.185529      241.92785    59.8682797
          430763         1273.75822     364.258421    1479.83294

The top line is production and the bottom is test.

This result supported the development team’s assertion that test was slower than production. The 32 queries averaged about 3.5 times longer run times in test than in production. Also, the time spent on I/O was about 25 times worse. I am not sure why the I/O time exceeded the elapsed time on test. I guess it has something to do with how Oracle measures I/O time. But clearly on average these 32 queries are much slower on test and I/O time probably caused most of the run time difference.

After noticing this big difference between test and production I decided to get these same sorts of performance metrics for each signature to see if certain ones were worse than others. The query looked like this:

sum(executions_delta) total_executions,
from DBA_HIST_SQLSTAT ss,common_sigs cs
sum(executions_delta) > 0
group by
order by

I put together the outputs from running this query on test and production and lined the result up like this:

------------------------ ------------------ ------------------
      575231776450247964         20268.6719         16659.4585
      944718698451269965         727534.558          3456111.6 *
     1037345866341698119         6640.87641         8859.53518
     1080231657361448615         3611.37698         4823.62857
     2879196232471320459         95723.5569         739287.601 *
     2895012443099075884         687272.949         724081.946
     3371400666194280661         1532797.66         761762.181
     4156520416999188213         109238.997         213658.722
     4634693999459450255          4923.8897         4720.16455
     5447362809447709021         2875.37308          2659.5754
     5698160695928381586         17139.6304         16559.1932
     6260911340920427003         290069.674         421058.874 *
     7412302135920006997         20039.0452         18951.6357
     7723300319489155163         18045.9756         19573.4784
     9153380962342466451         1661586.53         1530076.01
     9196714121881881832         5.48003488         5.13169472
     9347242065129163091         4360835.92         4581093.93
    11140980711532357629         3042320.88         5048356.99
    11160330134321800286         6868746.78         6160556.38
    12212345436143033196          5189.7972         5031.30811
    12776764566159396624         139150.231         614207.784  *
    12936428121692179551         3563.64537         3436.59365
    13637202277555795727          7360.0632         6410.02772
    14067042856362022182         859.732015         771.041714
    14256464986207527479         51.4042938         48.9237251
    14707568089762185958         627.586095          414.14762
    15001584593434987669         1287629.02         1122151.35
    15437049687902878835         96014.9782         996974.876  *
    16425440090840528197         48013.8912         50799.6184
    16778386062441486289         29459.0089         26845.8327
    17620933630628481201         51199.0511         111785.525  *
    18410003796880256802         581563.611         602866.609

I put an asterisk (*) beside the six queries that were much worse on test than production. I decided to focus on these six to get to the bottom of the reason between the difference. Note that many of the 32 queries ran about the same on test as prod so it really isn’t the case that everything was slow on test.

Now that I had identified the 6 queries I wanted to look at what they were spending their time on including both CPU and wait events. I used the following query to use ASH to get a profile of the time spent by these queries on both databases:

when 'WAITING' then event
(count(*)*10) seconds
order by seconds desc;

The profile looked like this in test:

------------------------ -------
db file parallel read     207450
ON CPU                    141010
db file sequential read    62990
direct path read           36980
direct path read temp      29240
direct path write temp     23110

The profile looked like this in production:

------------------------ -------
ON CPU                    433260
PX qref latch              64200
db file parallel read      35730
db file sequential read    14360
direct path read           12750
direct path write temp     12000

So, I/O waits dominate the time on test but not production. Since db file parallel read and db file sequential read were the top I/O waits for these 6 queries I used ash to see which of the 6 spent the most time on these waits.

db file parallel read:

  2  sql_id,
  3  (count(*)*10) seconds
  5  where
  7  ('944718698451269965',
  8  '2879196232471320459',
  9  '6260911340920427003',
 10  '12776764566159396624',
 11  '15437049687902878835',
 12  '17620933630628481201') and
 13  event='db file parallel read'
 14  group by sql_id
 15  order by seconds desc;

SQL_ID           SECONDS
------------- ----------
ak2wk2sjwnd34     159020
95b6t1sp7y40y      37030
brkfcwv1mqsas      11370
7rdc79drfp28a         30

db file sequential read:

  2  sql_id,
  3  (count(*)*10) seconds
  5  where
  7  ('944718698451269965',
  8  '2879196232471320459',
  9  '6260911340920427003',
 10  '12776764566159396624',
 11  '15437049687902878835',
 12  '17620933630628481201') and
 13  event='db file sequential read'
 14  group by sql_id
 15  order by seconds desc;

SQL_ID           SECONDS
------------- ----------
95b6t1sp7y40y      26840
ak2wk2sjwnd34      22550
6h0km9j5bp69t      13300
brkfcwv1mqsas        170
7rdc79drfp28a        130

Two queries stood out at the top waiters on these two events: 95b6t1sp7y40y and ak2wk2sjwnd34. Then I just ran my normal sqlstat query for both sql_ids for both test and production to find out when they last ran. Here is what the query looks like for ak2wk2sjwnd34:

select ss.sql_id,
ELAPSED_TIME_DELTA/(executions_delta*1000) "Elapsed Average ms",
CPU_TIME_DELTA/(executions_delta*1000) "CPU Average ms",
IOWAIT_DELTA/(executions_delta*1000) "IO Average ms",
CLWAIT_DELTA/(executions_delta*1000) "Cluster Average ms",
APWAIT_DELTA/(executions_delta*1000) "Application Average ms",
CCWAIT_DELTA/(executions_delta*1000) "Concurrency Average ms",
BUFFER_GETS_DELTA/executions_delta "Average buffer gets",
DISK_READS_DELTA/executions_delta "Average disk reads",
ROWS_PROCESSED_DELTA/executions_delta "Average rows processed"
where ss.sql_id = 'ak2wk2sjwnd34'
and ss.snap_id=sn.snap_id
and executions_delta > 0
order by ss.snap_id,ss.sql_id;

I found two time periods where both of these queries were recently run on both test and production and got an AWR report for each time period to compare them.

Here are a couple of pieces of the AWR report for the test database:



Here are similar pieces for the production database:

top5 foreground elapsed

What really stood out to me was that the wait events were so different. In production the db file parallel read waits averaged around 1 millisecond and the db file sequential reads averaged under 1 ms. On test they were 26 and 5 milliseconds, respectively. The elapsed times for sql_ids 95b6t1sp7y40y and ak2wk2sjwnd34 were considerably longer in test.

This is as far as my investigation went. I know that the slowdown is most pronounced on the two queries and I know that their I/O waits correspond to the two wait events. I am still trying to find a way to bring the I/O times down on our test database so that it more closely matches production. But at least I have a more narrow focus with the two top queries and the two wait events.


Categories: DBA Blogs

Jonathan Lewis

Tue, 2016-04-19 18:09

I am finally getting around to finishing my four-part blog series on people who have had the most influence on my Oracle performance tuning work. The previous three people were Craig ShallahamerDon Burleson, and Cary Millsap. The last person is Jonathan Lewis. These four people, listed and blogged about in chronological order, had the most influence on my understanding of how to do Oracle database performance tuning. There are many other great people out there and I am sure that other DBAs would produce their own, different, list of people who influenced them. But this list reflects my journey through my Oracle database career and the issues that I ran into and the experiences that I had. I ran into Jonathan Lewis’ work only after years of struggling with query tuning and getting advice from others. I ran into his material right around the time that I was beginning to learn about how the Oracle optimizer worked and some of its limits. Jonathan was a critical next step in my understanding of how Oracle’s optimizer worked and why it sometimes failed to pick the most efficient way to run a query.

Jonathan has produced many helpful tuning resources including his blog, his participation in online forums, and his talks at user group conferences, but the first and most profound way he taught me about Oracle performance tuning was through his query tuning book Cost-Based Oracle Fundamentals. It’s $30 on Amazon and that is an incredibly small amount of money to pay compared to the value of the material inside the book. I had spent many hours over several years trying to understand why the Oracle optimizer some times choses the wrong way to run a query. In many cases the fast way to run something was clear to me and the optimizer’s choices left me stumped. The book helped me better understand how the Oracle optimizer chooses what it thinks is the best execution plan. Jonathan’s book describes the different parts of a plan – join types, access methods, etc. – and how the optimizer assigns a cost to the different pieces of a plan. The optimizer chooses the plan with the least cost, but if some mistake causes the optimizer to calculate an unrealistic cost then it might choose a poor plan. Understanding why the optimizer would choose a slow plan helped me understand how to resolve performance issues or prevent them from happening, a very valuable skill.

There is a lot more I could say about what I got from Jonathan Lewis’ book including just observing how he operated. Jonathan filled his book with examples which show concepts that he was teaching. I think that I have emulated the kind of building of test scripts that you see throughout his book and on his blog and community forums. I think I have emulated not only Jonathan’s approach but the approaches of all four of the people who I have spotlighted in this series. Each have provided me with profoundly helpful technical information that has helped me in my career. But they have also provided me with a pattern of what an Oracle performance tuning practitioner looks like. What kind of things do they do? To this point in my career I have found the Oracle performance tuning part of my job to be the most challenging and interesting and probably the most valuable to my employers. Jonathan Lewis and the three others in this four-part series have been instrumental in propelling me along this path and I am very appreciative.


Categories: DBA Blogs

Log file parallel write wait graph

Thu, 2016-03-31 10:50

I got a chance to use my onewait Python based graph to help with a performance problem. I’m looking at slow write time from the log writer on Thursday mornings. Here is the graph with the database name erased:


We are still trying to track down the source of the problem but there seems to be a backup on another system that runs at times that correspond to the spike in log file parallel write wait times. The nice thing about this graph is that it shows you activity on the top and average wait time on the bottom so you can see if the increased wait time corresponds to a spike in activity. In this case there does not seem to be any increase in activity on the problematic database.  But that makes sense if the real problem is contention by a backup on another system.

Anyway, my Python graphs are far from perfect but still helpful in this case.


Categories: DBA Blogs

Python DBA Graphs Github Repository

Tue, 2016-03-29 17:40

I decided to get rid of the Github repository that I had experimented with and to create a new one. The old one had a dump of all my SQL scripts but without any documentation. But, I have updated my Python graphing scripts a bit at a time and have had some recent value from these scripts in my Oracle database tuning work. So, I created a Github repository called PythonDBAGraphs. I think it will be more valuable to have a repository that is more focused and is being actively updated and documented.

It is still very simple but I have gotten real value from the two graphs that are included.


Categories: DBA Blogs

Another SQL Profile to the rescue!

Mon, 2016-03-28 18:57

We have had problems with set of databases over the past few weeks. Our team does not support these databases, but my director asked me to help. These are Windows 64 bit Oracle databases running on Windows 2008. The incident reports said that the systems stop working and that the main symptom was that the oracle.exe process uses all the CPU. They were bouncing the database server when they saw this behavior and it took about 30 minutes after the bounce for the CPU to go back down to normal. A Windows server colleague told me that at some point in the past a new version of virus software had apparently caused high CPU from the oracle.exe process.

At first I looked for some known bugs related to high CPU and virus checkers without much success. Then I got the idea of just checking for query performance. After all, a poorly performing query can eat up a lot of CPU. These Windows boxes only have 2 cores so it would not take many concurrently running high CPU queries to max it out. So, I got an AWR report covering the last hour of a recent incident. This was the top SQL:


The top query, sql id 27d8x8p6139y6, stood out as very inefficient and all CPU. It seemed clear to me from this listing that the 2 core box had a heavy load and a lot of waiting for CPU queuing. %IO was zero but %CPU was only 31%. Most likely the rest was CPU queue time.

I also looked at my sqlstat report to see which plans 27d8x8p6139y6 had used over time.

--------------- --------------------- ---------- -----------
     3067874494 07-MAR-16 09.00.50 PM        287  948.102286
     3067874494 07-MAR-16 10.00.03 PM        292  1021.68191
     3067874494 07-MAR-16 11.00.18 PM        244  1214.96161
     3067874494 08-MAR-16 12.00.32 AM        276  1306.16222
     3067874494 08-MAR-16 01.00.45 AM        183  1491.31307
      467860697 08-MAR-16 01.00.45 AM        125      .31948
      467860697 08-MAR-16 02.00.59 AM        285  .234073684
      467860697 08-MAR-16 03.00.12 AM        279  .214354839
      467860697 08-MAR-16 04.00.25 AM        246   .17147561
      467860697 08-MAR-16 05.00.39 AM         18        .192
     2868766721 13-MAR-16 06.00.55 PM         89    159259.9
     3067874494 13-MAR-16 06.00.55 PM          8  854.384125
     2868766721 13-MAR-16 07.00.50 PM         70  1331837.56

Plan 2868766721 seemed terrible but plan 467860697 seemed great.

Our group doesn’t support these databases so I am not going to dig into how the application gathers statistics, what indexes it uses, or how the vendor designed the application. But, it seems possible that forcing the good plan with a SQL Profile could resolve this issue without having any access to the application or understanding of its design.

But, before plunging headlong into the use of a SQL Profile I looked at the plan and the SQL text.  I have edited these to hide any proprietary details:

         TAB_MYTABLE1 PIR_T,
         AND A.ASSIGNED_TO = :B1
         AND A.ACTIVE_FL = 1
         AND T.COMPLETE_FL = 0
         AND T.SHORTED_FL = 0
         AND ((PIR_A.FLOATING_PIR_FL = 1 
               AND PIR_T.COMPLETE_FL = 1)

The key thing I noticed is that there was only one bind variable. The innermost part of the good plan uses an index on the column that the query equates with the bind variable. The rest of the plan is a nice nested loops plan with range and unique index scans. I see plans in this format in OLTP queries where you are looking up small numbers of rows using an index and join to related tables.

Id | Operation                        | Name                     
 0 | SELECT STATEMENT                 |                          
 1 |  SORT ORDER BY                   |                          
 2 |   NESTED LOOPS                   |                          
 3 |    NESTED LOOPS                  |                          
 4 |     NESTED LOOPS                 |                          
 5 |      NESTED LOOPS                |                          
 7 |        INDEX RANGE SCAN          | AK_MYTABLELNG_BY_USER    
11 |       INDEX UNIQUE SCAN          | PK_MYTABLE1              
12 |     INDEX UNIQUE SCAN            | PK_MYTABLELNG            

The bad plan had a gross Cartesian merge join:

Plan hash value: 2868766721

Id | Operation                       | Name                     
 0 | SELECT STATEMENT                |                          
 1 |  NESTED LOOPS                   |                          
 2 |   NESTED LOOPS                  |                          
 3 |    MERGE JOIN CARTESIAN         |                          
 5 |      INDEX FULL SCAN            | PK_MYTABLE1              
 6 |     BUFFER SORT                 |                          
 8 |       INDEX RANGE SCAN          | AK_MYTABLELNG_BY_USER    
12 |    INDEX RANGE SCAN             | AK_MYTABLELNG_BY_USER    

Reviewing the SQL made me believe that there was a good chance that a SQL Profile forcing the good plan would resolve the issue. Sure, there could be some weird combination of data and bind variable values that make the bad plan the better one. But, given that this was a simple transactional application it seems most likely that the straightforward nested loops with index on the only bind variable plan would be best.

We used the SQL Profile to force these plans on four servers and so far the SQL Profile has resolved the issues. I’m not saying that forcing a plan using a SQL Profile is the only or even best way to resolve query performance issues. But, this was a good example of where a SQL Profile makes sense. If modifying the application, statistics, parameters, and schema is not possible then a SQL Profile can come to your rescue in a heartbeat.


Categories: DBA Blogs

Math Resources

Thu, 2016-03-17 17:51

I feel like I have not been posting very much on this blog lately. I have been focused on things outside of Oracle performance so I haven’t had a lot of new scripts to post.  I have been quietly updating my Python source code on GitHub so check that out. I have spent a lot of time educating myself in various ways including through the leadership and communication training program that comes from Toastmasters. My new job title is “Technical Architect” which is a form of technical leadership so I’m trying to expand myself beyond being an Oracle database administrator that specializes in performance tuning.

In addition to developing my leadership and communication skills I have gotten into a general computer science self-education kick. I took two introductory C.S. classes on edX. I also read a book on Linux hacking and a book on computer history. I was thinking of buying one of the Donald Knuth books or going through MIT’s free online algorithms class class 6.006. I have a computer science degree and spent two years in C.S. graduate school but that was a long time ago. It is kind of fun to refresh my memory and catch up with the latest trends. But the catch is that both the Knuth book and MIT’s 6.006 class require math that I either never learned or have forgotten. So, I am working my way through some math resources that I wanted to share with those who read this blog.

The first thing I did was to buy a computer math book, called Concrete Mathematics,  that seemed to cover the needed material. Reviews on recommended this book as good background for the Knuth series and one of the Oracle performance experts that I follow on Twitter recommended it for similar reasons. But, after finishing my second edX class I began exploring the MIT OCW math class that was a prerequisite to MIT’s 6.006 algorithms class. MIT calls the math class 6.042J and I am working through the Fall 2010 version of the class. There is a lot of overlap between the class and the book but they are not a perfect match. The book has some more difficult to follow material than the class. It is probably more advanced.  The class covers some topics, namely graph theory, that the book does not.  The free online class has some very good lecture videos by a top MIT professorTom Leighton. I even had my wife and daughters sit down and watch his first lecture with me on our family television for fun on my birthday.

The book led me to a great free math resource called Maxima. Maxima has all kinds of great math built into it such as solving equations, factoring integers, etc. Plus, it is free. There are other similar and I think more popular programs that are not free but for my use it was great to simply download Maxima and have its functionality at my fingertips.

The last resource that I wanted to mention is the Mathematics section of Stack Exchange. It is a pretty structured online forum with a question and answer format. It is helpful to me since I am going through 6.042J without a professor or teaching assistant to answer my questions. The people on math stack exchange are very helpful if you at least try to follow the etiquette for their forum. For example, they have an easy to use way to format math formulas in your questions and answers and the users of the forum expect you to use it. But it isn’t hard. I had one question from the Concrete Math book where I couldn’t understand the answer key in the back. I asked about it on stack exchange and got a great answer in no time.

Anyway, maybe all of this math and computer science study is a departure from my bread and butter Oracle database work and performance tuning. But the free online resources like the OCW web site, the Maxima program, and the stack exchange forum along with the book that I paid for are a great set of resources. I have already used some of the concepts that I have learned about number theory and its application to RSA encryption. But, at the same time I am enjoying studying these things and mostly see it as something fun to do in my spare time. (I’m weird I know.)

So, I have written this blog post to share the math related things that I am studying and using to those who might benefit from them. I am not a math expert, but I am getting a lot out of these materials. I hope that others find these resources as enjoyable and educational as I have.


Categories: DBA Blogs

General troubleshooting lessons from recent Delphix issue

Fri, 2016-02-05 11:25

Delphix support helped me resolve an issue yesterday and the experience gave me the idea of writing this post about several general computer issue troubleshooting tips that I have learned down through the years. Never mind that I ignored these lessons during this particular problem. This is more of a “do as I say” and not a “do as I do” story.  Actually, some times I remember these lessons. I didn’t do so well this week. But the several mistakes that I made resolving this recent Delphix issue motivate me to write this post and if nothing else remind myself of the lessons I’ve learned in the past about how to resolve a computer problem.

Don’t panic!

I’m reminded of the friendly advice on the cover of the Hitchhiker’s Guide to the Galaxy: “Don’t panic!”. So, yesterday it was 4:30 pm. I had rebooted the Delphix virtual machine and then in a panic had the Unix team reboot the HP Unix target server. But, still I could not bring up any of the Delphix VDBs.  We had people coming over to our house for dinner that night and I was starting to worry that I would be working on this issue all night. I ended up getting out of the office by 5:30 pm and had a great dinner with friends. What was I so stressed about? Even the times that I have been up all night it didn’t kill me. Usually the all night issues lead to me learning things anyway.

Trust support

The primary mistake that I made was to get my mind fixed on a solution to the problem instead of working with Delphix support and trusting them to guide us to the solution. We had a number of system issues due to a recent network issue and I got my mind set on the idea that my Delphix issue was due to some network hangup. I feel sorry for our network team because it seems like the first thought people have any time there is some issue is that it is a “network issue”. I should know better. How many times have I been working on issues when everyone says it is a “database issue” and I’m annoyed because I know that the issue is somewhere else and they are not believing me when I point to things outside the database. Anyway, I opened a case with Delphix on Monday when I couldn’t get a VDB to come down. It just hung for 5 minutes until it gave me an error. I assumed that it was a network hangup and got fixated on rebooting the Delphix VM. Ack! Ultimately, I ended up working with two helpful and capable people in Delphix support and they resolved the issue which was not what I thought at all. There are times to disagree with support and push for your own solution but I did this too early in this case and I was dead wrong.

Keep it simple

I’ve heard people refer to Occam’s razor which I translate in computer terms to mean “look for simple problems first”. Instead of fixing my mind on some vague network issue where the hardware is not working properly, how about assuming that all the hardware and software is working normally and then thinking about what problems might cause my symptoms? I can’t remember how many times this has bit me. There is almost always some simple explanation.  In this case I had made a change to a Unix shell script that runs when someone logs in as the oracle user. This caused Delphix to no longer be able to do anything with the VDBs on that server. Oops! It was a simple blunder, no big deal. But I’m kicking myself for not first thinking about a simple problem like a script change instead of focusing on something more exotic.

What changed?

I found myself saying the same dumb thing that I’ve heard people say to me all the time: nothing changed. In this case I said something like “this has worked fine for 3 years now and nothing has changed”. The long-suffering and patient Delphix support folks never called me on this, but I was dead wrong. Something had to have changed for something that was working to stop working. I should have spent time looking at the various parts of our Delphix setup to see if anything had changed before I contacted support. All I had to do was see the timestamp on our login script and I would see that something had recently changed.

Understand how it all works

I think my Delphix skills are a little rusty. We just started a new expansion project to add new database sources to Delphix. It has been a couple of years since I’ve done any heavy configuration and trouble shooting. But I used to have a better feel for how all the pieces fit together. I should have thought about what must have gone on behind the scenes when I asked Delphix to stop a VDB and it hung for 5 minutes. What steps was it doing? Where in the process could the breakdown be occurring? Delphix support did follow this type of reasoning to find the issue. They manually tried some of the steps that the Delphix software would do automatically until they found the problem. If I stopped to think about the pieces of the process I could have done the same. This has been a powerful approach to solving problems all through my career. I think about resolving PeopleSoft issues. It just helps to understand how things work. For example, if you understand how the PeopleSoft login process works you can debug login issues by checking each step of the process for possible issues. The same is true for Oracle logins from clients. In general, the more you understand all the pieces of a computer system, down to the transistors on the chips, the better chance you have of visualizing where the problem might be.

Well, I can’t think of any other pearls of wisdom from this experience but I thought I would write these down while it was on my mind. Plus, I go on call Monday morning so I need to keep these in mind as I resolve any upcoming issues. Thanks to Delphix support for their good work on this issue.

Categories: DBA Blogs