You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafodion.apache.org by "Alice Chen (JIRA)" <ji...@apache.org> on 2015/07/22 20:18:26 UTC
[jira] [Created] (TRAFODION-851) LP Bug: 1405235 - SB_HISTOGRAMS
table not found is incorrectly reported
Alice Chen created TRAFODION-851:
------------------------------------
Summary: LP Bug: 1405235 - SB_HISTOGRAMS table not found is incorrectly reported
Key: TRAFODION-851
URL: https://issues.apache.org/jira/browse/TRAFODION-851
Project: Apache Trafodion
Issue Type: Bug
Components: sql-security
Reporter: Roberta Marton
Assignee: Barry Fritchman
Priority: Critical
Fix For: 1.0 (pre-incubation)
There is a set of histogram tables created per schema. When update statistics is executed and the tables do not exist, update statistics creates them before continuing.
There is a regression where an error is returned indicating that the histogram tables do not exist when in fact they have been created.
To re-created:
*In one session, perform:
Create shared schema t132sch;
set schema t132sch;
get tables;
create table t132t1 (c1 int, c2 int);
create table t132t2 (c1 int, c2 int) attribute by sql_user1;
insert into t132t1 values (1,1), (2,2), (3,3),(4,4), (5,5),(6,6), (7,7),(8,8);
insert into t132t2 values (1,1), (2,2), (3,3),(4,4), (5,5),(6,6), (7,7),(8,8);
get tables;
select count(*) from t132t1;
select count(*) from t132t2;
*Log off and create a new sqlci session and do:
set schema t132sch;
update statistics for table t132t1 on every column;
*** ERROR[9200] UPDATE STATISTICS for table TRAFODION.T132SCH.T132T1 encountered an error (4082) from statement .
*** ERROR[4082] Object TRAFODION.T132SCH.SB_HISTOGRAMS does not exist or is inaccessible.
*** ERROR[8822] The statement was not prepared.
--- SQL operation failed with errors.
When putting the arkcmp that performs the update statistics into debug:
The code to create the histogram tables is being executed and tables are created. I can go to a different sqlci session and see the tables.
During stats collection, I am getting an internal error from the cli (8898) in Descriptor::setDescItem
The assertion in Descriptor::setDescItem is failing because the rowsetSize is 0.
#0 ComDiagsArea::insertNewError (this=0x7fffe9821830) at ../export/ComDiags.cpp:2807
#1 0x00007ffff6387d89 in ComDiagsArea::acceptNewCondition (this=0x7fffe9821830)
at ../export/ComDiags.cpp:2824
#2 0x00007ffff6ed41c9 in operator<< (d=..., dgObj=...) at ../common/DgBaseType.cpp:139
#3 0x00007ffff762df34 in Descriptor::setDescItem (this=0x7fffe8dec2f0, entry=0,
what_to_set=36, numeric_value=140737488313360, string_value=0x0, info_desc=0x0,
info_desc_index=0) at ../cli/Descriptor.cpp:2702
#4 0x00007ffff75e5724 in SQLCLI_SetRowsetDescPointers(CliGlobals *, SQLDESC_ID *, Lng32, Lng32 *, Lng32, Lng32, typedef __va_list_tag __va_list_tag *, SQLCLI_QUAD_FIELDS *) (
cliGlobals=0xb9ed80, desc_id=0x7fffdbf6b9d0, rowset_size=0,
rowset_status_ptr=0x7fffffff5c10, starting_entry=1, num_quadruple_fields=3, ap=0x0,
quad_fields=0x7fffdcbab1b0) at ../cli/Cli.cpp:1857
#5 0x00007ffff767cbf3 in SQL_EXEC_SETROWSETDESCPOINTERS (sql_descriptor=0x7fffdbf6b9d0,
rowset_size=0, rowset_status_ptr=0x7fffffff5c10, starting_entry=1,
num_quadruple_fields=3, quad_fields=0x7fffdcbab1b0) at ../cli/CliExtern.cpp:5602
#6 0x00007ffff5c42d19 in HSCursor::setRowsetPointers (this=0x7fffffff5df0, group=0x0,
maxRows=0) at ../ustat/hs_cli.cpp:2880
#7 0x00007ffff5c7e1d1 in HSGlobalsClass::readColumnsIntoMem (this=0x7fffffff6750,
cursor=0x7fffffff5df0, rows=8) at ../ustat/hs_globals.cpp:10626
#8 0x00007ffff5c6bb15 in HSGlobalsClass::CollectStatistics (this=0x7fffffff6750)
at ../ustat/hs_globals.cpp:5404
#9 0x00007ffff5cdc53d in UpdateStats (
input=0x7fffdcbe1808 "update statistics for table t132t1 on every column;",
requestedByCompiler=0) at ../ustat/hs_update.cpp:421
The code continues processing and error 8898 comes again during HSCursor destruction:
ComDiagsArea::insertNewError (this=0x7fffe9821830) at ../export/ComDiags.cpp:2804
#1 0x00007ffff6387d89 in ComDiagsArea::acceptNewCondition (this=0x7fffe9821830)
at ../export/ComDiags.cpp:2824
#2 0x00007ffff6388a98 in ComDiagsArea::mergeAfter (this=0x7fffe9821830, source=...)
at ../export/ComDiags.cpp:3186
#3 0x00007ffff3478f7b in ex_root_tcb::cancel (this=0x7fffe8da91a0, glob=0x7fffe8dcc610,
diagsArea=@0x7fffffff5a90, getQueueDiags=0) at ../executor/ex_root.cpp:2381
#4 0x00007ffff76576f6 in CliStatement::close (this=0x7fffe8da61e8, diagsArea=...,
inRollback=0) at ../cli/Statement.cpp:1146
#5 0x00007ffff75e9abf in SQLCLI_PerformTasks(CliGlobals *, ULng32, SQLSTMT_ID *, SQLDESC_ID *, SQLDESC_ID *, Lng32, Lng32, typedef __va_list_tag __va_list_tag *, SQLCLI_PTR_PAIRS *, SQLCLI_PTR_PAIRS *) (cliGlobals=0xb9ed80, tasks=776, statement_id=0x7fffdcb93728,
input_descriptor=0x0, output_descriptor=0x0, num_input_ptr_pairs=0,
num_output_ptr_pairs=0, ap=0x0, input_ptr_pairs=0x0, output_ptr_pairs=0x0)
at ../cli/Cli.cpp:3451
#6 0x00007ffff75e9d03 in SQLCLI_CloseStmt (cliGlobals=0xb9ed80,
statement_id=0x7fffdcb93728) at ../cli/Cli.cpp:3504
#7 0x00007ffff76767e2 in SQL_EXEC_CloseStmt (statement_id=0x7fffdcb93728)
at ../cli/CliExtern.cpp:1404
#8 0x00007ffff5c4114d in HSCursor::~HSCursor (this=0x7fffffff5df0,
__in_chrg=<value optimized out>) at ../ustat/hs_cli.cpp:2573
#9 0x00007ffff5c6bfbc in HSGlobalsClass::CollectStatistics (this=0x7fffffff6750)
at ../ustat/hs_globals.cpp:5397
#10 0x00007ffff5cdc53d in UpdateStats (
input=0x7fffdcbe1808 "update statistics for table t132t1 on every column;",
requestedByCompiler=0) at ../ustat/hs_update.cpp:421
Then flushStatistics is called and error 4082 is returned when trying to read histograms table. The error is coming back from the second arkcmp. I didn’t put the second arkcmp into debug so don’t know why it is failing to find the table. As I mentioned earlier, I can bring up a sqlci session at the same time and see these tables.
0x00007ffff767bce4 in SQL_EXEC_Prepare2 (statement_id=0x7fffdbf71460,
sql_source=0x7fffdbf71498, gencode_ptr=0x0, gencode_len=0, ret_gencode_len=0x0,
query_cost_info=0x7fffffff60b0, comp_stats_info=0x7fffffff4fb0, uniqueStmtId=0x0,
uniqueStmtIdLen=0x0, flags=0) at ../cli/CliExtern.cpp:4958
#23 0x00007ffff5c4fa0d in HSCursor::prepareQuery (this=0x7fffffff6260,
cliStr=0x17be068 "SELECT HISTOGRAM_ID, COL_POSITION, COLUMN_NUMBER, COLCOUNT, cast(READ_TIME as char(19)), REASON FROM TRAFODION.T132SCH.SB_HISTOGRAMS WHERE TABLE_UID = 36082714901494630 ORDER BY TABLE_UID, HISTOGRAM_I"..., numParams=0, numResults=6)
at ../ustat/hs_cli.cpp:6052
#24 0x00007ffff5c74945 in HSGlobalsClass::groupListFromTable (this=0x7fffffff6750,
groupList=@0x7fffffff6518, skipEmpty=0, exclusive=1) at ../ustat/hs_globals.cpp:7673
#25 0x00007ffff5c66e8e in HSGlobalsClass::MakeAllHistid (this=0x7fffffff6750)
at ../ustat/hs_globals.cpp:4023
#26 0x00007ffff5c726e0 in HSGlobalsClass::FlushStatistics (this=0x7fffffff6750,
statsWritten=@0x7fffffff6adc) at ../ustat/hs_globals.cpp:7179
#27 0x00007ffff5cdc774 in UpdateStats (
input=0x7fffdcbe1808 "update statistics for table t132t1 on every column;",
requestedByCompiler=0) at ../ustat/hs_update.cpp:444
Eventually 4082 is returned along with 9200 and 8822.
If I rerun the update statistics a second time in the same session, I get:
>>update statistics for table t132t1 on every column;
*** ERROR[1389] Object T132T1 does not exist in Trafodion.
*** ERROR[4082] Object TRAFODION.T132SCH.T132T1 does not exist or is inaccessible.
--- SQL operation failed with errors.
And if I rerun it a third time, it succeeds.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)