[
https://issues.apache.org/jira/browse/TRAFODION-2617?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16126179#comment-16126179
]
ASF GitHub Bot commented on TRAFODION-2617:
-------------------------------------------
GitHub user DaveBirdsall opened a pull request:
https://github.com/apache/incubator-trafodion/pull/1206
[TRAFODION-2617] Optionally push estimate row count logic into HBase
coprocessor
When HBase encryption is used, client-side logic to estimate row count in
encrypted Trafodion tables fails because the Trafodion ID lacks permission to
read the HBase key store file. This set of changes pushes the row count logic
down into a coprocessor, so the logic can run under the HBase ID.
The feature is turned on via CQD HBASE_ESTIMATE_ROW_COUNT_VIA_COPROCESSOR
'ON'. The default is set to 'OFF'. There is a small performance overhead when
using the coprocessor, so it is best to leave this set to 'OFF' unless HBase
encryption is employed.
Note that this change is not sufficient to support encrypted Trafodion
tables using HBase encryption. In order to specify at create time that a
Trafodion table should be encrypted, one needs an extension to the DDL
HBASE_OPTIONS (e.g., "ENCRYPTION = 'AES'"). This has not yet been implemented.
Nevertheless, one can still test these changes on a Trafodion instance by
using the HBase shell. After creating a table within Trafodion, one can use the
'alter' command in the HBase shell to turn encryption on. For example: alter
"TRAFODION.SEABASE.T1",NAME=>'#1',ENCRYPTION=>'AES'. That will come in a future
JIRA.
Other pitfalls to watch out for: Configure your instance for HBase
encryption as described in the HBase documentation. In addition, set the
hfile.format.version value to three. (In Trafodion, by default it is two.)
Oddly, HBase allows one to turn on encryption for format two tables, but it
does not work; you get file corruption errors.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/DaveBirdsall/incubator-trafodion
Trafodion2617a
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-trafodion/pull/1206.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #1206
----
commit 8b93ad50673a1be3ca9ae0e21eb00bb5627cc03a
Author: Dave Birdsall <[email protected]>
Date: 2017-08-14T18:17:53Z
[TRAFODION-2617] Optionally push estimate row count logic into HBase coproc
----
> Error 9252 during update statistics of an encrypted Trafodion table
> -------------------------------------------------------------------
>
> Key: TRAFODION-2617
> URL: https://issues.apache.org/jira/browse/TRAFODION-2617
> Project: Apache Trafodion
> Issue Type: Bug
> Components: sql-cmp
> Affects Versions: 2.1-incubating
> Environment: Any, HBase encryption is enabled for the table.
> Reporter: Hans Zeller
> Assignee: David Wayne Birdsall
>
> Anu tried an update statistics command for a table that is using HBase
> encryption. That failed with the following stack trace, as printed
> >>update statistics for table t on every column sample;
> ..
> *** ERROR[9252] Unable to get row count estimate: Error code 68, detail 4.
> Exception info (if any):
> Instead of showing the exception info printed to stdout, I'm showing the
> contents of the ulog file:
> UPDATE STATISTICS
> =====================================================================
> [Wed 17 May 2017 10:38:30 PM UTC] update statistics for table t on every
> column sample;
> [Wed 17 May 2017 10:38:30 PM UTC] :BEGIN UpdateStats()
> [Wed 17 May 2017 10:38:30 PM UTC] :| BEGIN Setup CQDs prior to parsing
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT QUERY_CACHE '0'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT CACHE_HISTOGRAMS 'OFF'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> USTAT_MODIFY_DEFAULT_UEC '0.05'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT OUTPUT_DATE_FORMAT
> 'ANSI'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> HIST_MISSING_STATS_WARNING_LEVEL '0'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> USTAT_AUTOMATION_INTERVAL '0'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> MV_ALLOW_SELECT_SYSTEM_ADDED_COLUMNS 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> HIST_ON_DEMAND_STATS_SIZE '0'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT ISOLATION_LEVEL 'READ
> COMMITTED'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> ALLOW_DML_ON_NONAUDITED_TABLE 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> MV_ALLOW_SELECT_SYSTEM_ADDED_COLUMNS 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> ALLOW_NULLABLE_UNIQUE_KEY_CONSTRAINT 'OFF'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> CAT_ERROR_ON_NOTNULL_STOREBY 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> WMS_CHILD_QUERY_MONITORING 'OFF'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT WMS_QUERY_MONITORING
> 'OFF'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> TRAF_TINYINT_RETURN_VALUES 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT TRAF_BOOLEAN_IO 'ON'
> [Wed 17 May 2017 10:38:30 PM UTC] CONTROL QUERY DEFAULT
> TRAF_LARGEINT_UNSIGNED_IO 'ON'
> [Wed 17 May 2017 10:38:31 PM UTC] CONTROL QUERY DEFAULT
> TRAF_ALLOW_RESERVED_COLNAMES 'ON'
> [Wed 17 May 2017 10:38:31 PM UTC] CONTROL QUERY DEFAULT TRAF_BLOB_AS_VARCHAR
> 'OFF'
> [Wed 17 May 2017 10:38:31 PM UTC] CONTROL QUERY DEFAULT TRAF_CLOB_AS_VARCHAR
> 'OFF'
> [Wed 17 May 2017 10:38:31 PM UTC] :| END Setup CQDs prior to parsing
> elapsed time (00:00:00.420)
> [Wed 17 May 2017 10:38:31 PM UTC] :| BEGIN Parse statement
> [Wed 17 May 2017 10:38:31 PM UTC] call HSHbaseTableDef::objExists
> [Wed 17 May 2017 10:38:31 PM UTC] naTbl_->objectUid() is 6001738912217799228
> [Wed 17 May 2017 10:38:31 PM UTC] CONTROL QUERY DEFAULT
> DISPLAY_DIVISION_BY_COLUMNS RESET
> [Wed 17 May 2017 10:38:31 PM UTC]
> CHECK SCHEMA VERSION FOR TABLE: XXXXXXXXXXXX
> [Wed 17 May 2017 10:38:31 PM UTC]
> UpdateStats: TABLE: XXXXXXXXXXXX; SCHEMA VERSION: 2600; AUTOMATION INTERVAL: 0
> [Wed 17 May 2017 10:38:31 PM UTC] KEY:
> (_SALT_,PATH_ID,NAME_ID)
> [Wed 17 May 2017 10:38:31 PM UTC] GroupExists: argument: colSet
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[0]: :_SALT_: 12
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[1]: :PATH_ID: 0
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[2]: :NAME_ID: 1
> [Wed 17 May 2017 10:38:31 PM UTC] KEY: (_SALT_,PATH_ID)
> [Wed 17 May 2017 10:38:31 PM UTC] GroupExists: argument: colSet
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[0]: :_SALT_: 12
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[1]: :PATH_ID: 0
> [Wed 17 May 2017 10:38:31 PM UTC] GroupExists: mgroup->colSet
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[0]: :_SALT_: 12
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[1]: :PATH_ID: 0
> [Wed 17 May 2017 10:38:31 PM UTC] colSet[2]: :NAME_ID: 1
> [Wed 17 May 2017 10:38:31 PM UTC] :| END Parse statement elapsed time
> (00:00:00.930)
> [Wed 17 May 2017 10:38:31 PM UTC]
> USTAT_CQDS_ALLOWED_FOR_SPAWNED_COMPILERS size of (0) is not acceptable
> [Wed 17 May 2017 10:38:31 PM UTC] :| BEGIN Initialize environment
> [Wed 17 May 2017 10:38:31 PM UTC] Creating histogram tables for schema
> TRAFODION.XXXXXXX on demand.
> [Wed 17 May 2017 10:38:31 PM UTC] :| | BEGIN Create histogram tables
> [Wed 17 May 2017 10:38:31 PM UTC] BEGIN WORK
> [Wed 17 May 2017 10:38:32 PM UTC] BEGINWORK(Create histogram tables.)
> [Wed 17 May 2017 10:38:32 PM UTC] Transaction started: 2017-05-17
> 22:38:32.007401
> [Wed 17 May 2017 10:38:33 PM UTC] :| | END Create histogram tables
> elapsed time (00:00:01.090)
> [Wed 17 May 2017 10:38:33 PM UTC] COMMIT WORK
> [Wed 17 May 2017 10:38:33 PM UTC] COMMITWORK()
> [Wed 17 May 2017 10:38:33 PM UTC] Transaction committed: 2017-05-17
> 22:38:33.099332
> [Wed 17 May 2017 10:38:33 PM UTC] :| | BEGIN getRowCount()
> [Wed 17 May 2017 10:38:33 PM UTC] :| | END getRowCount() elapsed time
> (00:00:00.065)
> [Wed 17 May 2017 10:38:33 PM UTC] currentRowCountIsEstimate_=1 from
> getRowCount()
> [Wed 17 May 2017 10:38:33 PM UTC] errorCode=68, breadCrumb=4
> [Wed 17 May 2017 10:38:33 PM UTC] JNI exception info:
> [Wed 17 May 2017 10:38:33 PM UTC]
> org.apache.hadoop.hbase.io.hfile.CorruptHFileException: Problem reading HFile
> Trailer from file
> hdfs://ip-172-31-65-71.ec2.internal:8020/apps/hbase/data/data/default/XXXXXXXXXX/00c6a0e9c39b98bd04f188647bd50253/#1/033b3f07b7c84725b5bc9e7aaf75eb54
> org.apache.hadoop.hbase.io.hfile.HFile.pickReaderVersion(HFile.java:481)
> org.apache.hadoop.hbase.io.hfile.HFile.createReader(HFile.java:524)
> org.trafodion.sql.HBaseClient.estimateRowCountBody(HBaseClient.java:1302)
> org.trafodion.sql.HBaseClient.estimateRowCount(HBaseClient.java:1207) Caused
> by
> java.lang.RuntimeException: java.lang.RuntimeException:
> java.io.FileNotFoundException: /etc/hbase/conf/hbase.jks (Permission denied)
> org.apache.hadoop.hbase.io.crypto.Encryption.getKeyProvider(Encryption.java:560)
> org.apache.hadoop.hbase.io.crypto.Encryption.getSecretKeyForSubject(Encryption.java:427)
> org.apache.hadoop.hbase.io.crypto.Encryption.decryptWithSubjectKey(Encryption.java:474)
> org.apache.hadoop.hbase.security.EncryptionUtil.getUnwrapKey(EncryptionUtil.java:129)
> org.apache.hadoop.hbase.security.EncryptionUtil.unwrapKey(EncryptionUtil.java:122)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.createHFileContext(HFileReaderV3.java:107)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV2.<init>(HFileReaderV2.java:130)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.<init>(HFileReaderV3.java:77)
> org.apache.hadoop.hbase.io.hfile.HFile.pickReaderVersion(HFile.java:471)
> org.apache.hadoop.hbase.io.hfile.HFile.createReader(HFile.java:524)
> org.trafodion.sql.HBaseClient.estimateRowCountBody(HBaseClient.java:1302)
> org.trafodion.sql.HBaseClient.estimateRowCount(HBaseClient.java:1207) Caused
> by
> java.lang.RuntimeException: java.io.FileNotFoundException:
> /etc/hbase/conf/hbase.jks (Permission denied)
> org.apache.hadoop.hbase.io.crypto.KeyStoreKeyProvider.init(KeyStoreKeyProvider.java:153)
> org.apache.hadoop.hbase.io.crypto.Encryption.getKeyProvider(Encryption.java:553)
> org.apache.hadoop.hbase.io.crypto.Encryption.getSecretKeyForSubject(Encryption.java:427)
> org.apache.hadoop.hbase.io.crypto.Encryption.decryptWithSubjectKey(Encryption.java:474)
> org.apache.hadoop.hbase.security.EncryptionUtil.getUnwrapKey(EncryptionUtil.java:129)
> org.apache.hadoop.hbase.security.EncryptionUtil.unwrapKey(EncryptionUtil.java:122)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.createHFileContext(HFileReaderV3.java:107)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV2.<init>(HFileReaderV2.java:130)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.<init>(HFileReaderV3.java:77)
> org.apache.hadoop.hbase.io.hfile.HFile.pickReaderVersion(HFile.java:471)
> org.apache.hadoop.hbase.io.hfile.HFile.createReader(HFile.java:524)
> org.trafodion.sql.HBaseClient.estimateRowCountBody(HBaseClient.java:1302)
> org.trafodion.sql.HBaseClient.estimateRowCount(HBaseClient.java:1207) Caused
> by
> java.io.FileNotFoundException: /etc/hbase/conf/hbase.jks (Permission denied)
> java.io.FileInputStream.open0(Native Method)
> java.io.FileInputStream.open(FileInputStream.java:195)
> java.io.FileInputStream.<init>(FileInputStream.java:138)
> org.apache.hadoop.hbase.io.crypto.KeyStoreKeyProvider.load(KeyStoreKeyProvider.java:124)
> org.apache.hadoop.hbase.io.crypto.KeyStoreKeyProvider.init(KeyStoreKeyProvider.java:147)
> org.apache.hadoop.hbase.io.crypto.Encryption.getKeyProvider(Encryption.java:553)
> org.apache.hadoop.hbase.io.crypto.Encryption.getSecretKeyForSubject(Encryption.java:427)
> org.apache.hadoop.hbase.io.crypto.Encryption.decryptWithSubjectKey(Encryption.java:474)
> org.apache.hadoop.hbase.security.EncryptionUtil.getUnwrapKey(EncryptionUtil.java:129)
> org.apache.hadoop.hbase.security.EncryptionUtil.unwrapKey(EncryptionUtil.java:122)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.createHFileContext(HFileReaderV3.java:107)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV2.<init>(HFileReaderV2.java:130)
> org.apache.hadoop.hbase.io.hfile.HFileReaderV3.<init>(HFileReaderV3.java:77)
> org.apache.hadoop.hbase.io.hfile.HFile.pickReaderVersion(HFile.java:471)
> org.apache.hadoop.hbase.io.hfile.HFile.createReader(HFile.java:524)
> org.trafodion.sql.HBaseClient.estimateRowCountBody(HBaseClient.java:1302)
> org.trafodion.sql.HBaseClient.estimateRowCount(HBaseClient.java:1207)
> [Wed 17 May 2017 10:38:33 PM UTC] :| END Initialize environment elapsed
> time (00:00:01.169)
> [Wed 17 May 2017 10:38:33 PM UTC] *** ERROR[-1] in hs_update:445
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
