[MDEV-12634] Uninitialised ROW_MERGE_RESERVE_SIZE bytes written to temporary file - Jira

Details

Type: Bug
Status: Closed (View Workflow)
Priority: Major
Resolution: Fixed
Affects Version/s: 10.1.7, 10.2.0, 10.3.0
Fix Version/s: 10.1.27, 10.2.9
Component/s: Storage Engine - InnoDB, Storage Engine - XtraDB
Labels:
None

Sprint:
10.1.24

Description

http://buildbot.askmonty.org/buildbot/builders/work-amd64-valgrind/builds/10024/steps/test/logs/stdio

innodb.innodb_bug53290 'innodb_plugin'   w1 [ fail ]  Found warnings/errors in server log file!

        Test ended at 2017-04-27 12:29:08

line

==31315== Thread 22:

==31315== Uninitialised byte(s) found during client check request

==31315==    at 0xC0F7216: row_merge_write_eof(unsigned char*, unsigned char*, int, unsigned long*, fil_space_crypt_t*, unsigned char*, unsigned long) (row0merge.cc:1360)

==31315==    by 0xC0F9CE2: row_merge_blocks(row_merge_dup_t const*, merge_file_t const*, unsigned char*, unsigned long*, unsigned long*, merge_file_t*, fil_space_crypt_t*, unsigned char*, unsigned long) (row0merge.cc:2328)

==31315==    by 0xC0FA449: row_merge(trx_t*, row_merge_dup_t const*, merge_file_t*, unsigned char*, int*, unsigned long*, unsigned long*, fil_space_crypt_t*, unsigned char*, unsigned long) (row0merge.cc:2484)

==31315==    by 0xC0FAB01: row_merge_sort(trx_t*, row_merge_dup_t const*, merge_file_t*, unsigned char*, int*, bool, float, float, fil_space_crypt_t*, unsigned char*, unsigned long) (row0merge.cc:2649)

==31315==    by 0xC0FE2F5: row_merge_build_indexes(trx_t*, dict_table_t*, dict_table_t*, bool, dict_index_t**, unsigned long const*, unsigned long, TABLE*, dtuple_t const*, unsigned long const*, unsigned long, ib_sequence_t&) (row0merge.cc:4230)

==31315==    by 0xC018846: ha_innodb::inplace_alter_table(TABLE*, Alter_inplace_info*) (handler0alter.cc:4249)

==31315==    by 0x6E90C6: handler::ha_inplace_alter_table(TABLE*, Alter_inplace_info*) (handler.h:3704)

==31315==    by 0x6E1256: mysql_inplace_alter_table(THD*, TABLE_LIST*, TABLE*, TABLE*, Alter_inplace_info*, enum_alter_inplace_result, MDL_request*, Alter_table_ctx*) (sql_table.cc:7133)

==31315==    by 0x6E5ACF: mysql_alter_table(THD*, char*, char*, HA_CREATE_INFO*, TABLE_LIST*, Alter_info*, unsigned int, st_order*, bool) (sql_table.cc:8945)

==31315==    by 0x75AFCD: Sql_cmd_alter_table::execute(THD*) (sql_alter.cc:325)

==31315==    by 0x6266AF: mysql_execute_command(THD*) (sql_parse.cc:5695)

==31315==    by 0x62A952: mysql_parse(THD*, char*, unsigned int, Parser_state*) (sql_parse.cc:7343)

==31315==  Address 0x18736000 is in a rw- anonymous segment

Attachments

Issue Links

is caused by

MDEV-8708 InnoDB temp file encryption

Closed

relates to

MDEV-13826 CREATE FULLTEXT INDEX on encrypted table fails

Closed

MDEV-13827 Add encryption tests for FULLTEXT INDEX

Closed

Activity

Descending order - Click to sort in ascending order

Jan Lindström (Inactive) added a comment - 2017-09-18 08:11

Correct, this effects only temporary files created and used by InnoDB internally.

Jan Lindström (Inactive) added a comment - 2017-09-18 08:11 Correct, this effects only temporary files created and used by InnoDB internally.

Ian Gilfillan added a comment - 2017-09-18 08:01

jplindst so this is regardless of the setting of encrypt_tmp_files, correct?

Ian Gilfillan added a comment - 2017-09-18 08:01 jplindst so this is regardless of the setting of encrypt_tmp_files, correct?

Jan Lindström (Inactive) added a comment - 2017-09-18 06:33

greenman Can you update the documentation to mention that in 10.1.27, 10.2.9,10.3.2 temporary files created by merge sort and row log are encrypted if innodb-encrypt-log=TRUE regardless is table encrypted or not.

Jan Lindström (Inactive) added a comment - 2017-09-18 06:33 greenman Can you update the documentation to mention that in 10.1.27, 10.2.9,10.3.2 temporary files created by merge sort and row log are encrypted if innodb-encrypt-log=TRUE regardless is table encrypted or not.

Marko Mäkelä added a comment - 2017-09-17 10:52

I pushed the merge to 10.2 and an adjustment to innodb-table-online that makes it use the merge sort files.

Marko Mäkelä added a comment - 2017-09-17 10:52 I pushed the merge to 10.2 and an adjustment to innodb-table-online that makes it use the merge sort files .

Marko Mäkelä added a comment - 2017-09-15 18:46 - edited

I would revert this instrumentation from 10.1 and instead, introduce an innodb_encrypt_log.combinations file and use it in the tests innodb.innodb-table-online and innodb.innodb-index-online.

	DBUG_EXECUTE_IF(

		"ib_merge_wait_after_read",

		os_thread_sleep(20000000););  /* 20 sec */

Those two tests use proper DEBUG_SYNC synchronization and do not introduce such unreasonable (and inherently nondeterministic) delays.
This is now building on bb-10.1-marko.

Marko Mäkelä added a comment - 2017-09-15 18:46 - edited I would revert this instrumentation from 10.1 and instead, introduce an innodb_encrypt_log.combinations file and use it in the tests innodb.innodb-table-online and innodb.innodb-index-online. DBUG_EXECUTE_IF( "ib_merge_wait_after_read" , os_thread_sleep(20000000);); /* 20 sec */ Those two tests use proper DEBUG_SYNC synchronization and do not introduce such unreasonable (and inherently nondeterministic) delays. This is now building on bb-10.1-marko .

View 11 older comments

People

Assignee:: Jan Lindström (Inactive)

Reporter:: Elena Stepanova

Votes:: 0 Vote for this issue

Watchers:: 4 Start watching this issue

Dates

Created:: 2017-04-29 00:46

Updated:: 2017-09-28 09:38

Resolved:: 2017-09-14 14:35

Git Integration

Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.

MariaDB Server