- 02 11月, 2016 3 次提交
-
-
由 Daniel Gustafsson 提交于
pgstat_write_statsfile() failed to write the resource queue statistics to the stat file which in turn makes the pg_stat_resqueues view empty. Patch by Github user LJoNe with testcase added by me
-
由 Haisheng Yuan 提交于
gporca has a set of banned API calls which needs to be allowed with the ALLOW_xxx macro in order for gpopt to compile. But it should be the library caller(GPDB/Orca)'s resposibility to take care of the function call. see discussions on greenplum-db/gpdb#1136 and https://groups.google.com/a/greenplum.org/forum/#!topic/gpdb-dev/Mcw6JPav6h4
-
由 foyzur 提交于
* Adding support for interrupt processing before reserving more vmem. * Process pending interrupts before reserving VMEM. * Adding guc to control vmem tracker checking for interrupts before reserving more vmem.
-
- 01 11月, 2016 8 次提交
-
-
由 brendanstephens 提交于
.psqlrc can create unexpected output and changes in formatting that don't play nice with parse_oids(). ``` psql database --pset footer -Atq -h localhost -p 5432 -U gpadmin -f /tmp/20161012232709/toolkit.sql {"relids": "573615536", "funcids": ""} Time: 2.973 ms ``` Generates an Exception: ``` Traceback (most recent call last): File "/usr/local/greenplum-db/./bin/minirepro", line 386, in <module> main() File "/usr/local/greenplum-db/./bin/minirepro", line 320, in main mr_query = parse_oids(cursor, json_str) File "/usr/local/greenplum-db/./bin/minirepro", line 151, in parse_oids result.relids = json.loads(json_oids)['relids'] File "/usr/local/greenplum-db/ext/python/lib/python2.6/json/__init__.py", line 307, in loads return _default_decoder.decode(s) File "/usr/local/greenplum-db/ext/python/lib/python2.6/json/decoder.py", line 322, in decode raise ValueError(errmsg("Extra data", s, end, len(s))) ValueError: Extra data: line 2 column 1 - line 3 column 1 (char 39 - 54) ```
-
由 Heikki Linnakangas 提交于
In many places where we had used a mixture of spaces and tabs for indentation, new versions of gcc complained about misleading indentation, because gcc doesn't know we're using tab width of 4. To fix, make the indentation consistent in all the places where gcc gave a warning. Would be nice to fix it all around, but that's a lot of work, so let's do it in a piecemeal fashion whenever we run into issues or need to modify a piece of code anyway. For some files, especially the GPDB-specific ones, I ran pgindent over the whole file. I used the pgindent from PostgreSQL master, which is slightly different from what was used back 8.3 days, but that's what I had easily available, and that's what we're heading to in the future anyway. In some cases, I didn't commit the pgindented result if there were funnily formatted code or comments that would need special treatment. For other places, I fixed the indentation locally, just enough to make the warnings go away. I also did a tiny bit of other trivial cleanup, that I happened to spot while working on this, although I tried to refrain from anything more extensive.
-
由 Adam Lee 提交于
Update storage types of cidr and inet in AOCO_Compression2, schema_topology and schema_topology_optimizer expected files. Related commit: 3e23b68dSigned-off-by: NPengzhou Tang <ptang@pivotal.io>
-
由 Heikki Linnakangas 提交于
Aggressive inlining can improve performance in hot spots, but when applied indiscriminately, it makes binaries larger for little gain. For lack of better evidence, let's trust the compiler defaults. This makes the build faster, and the binaries smaller. If we find out that these options are beneficial for some specific functions or source files, let's add them back to those special cases. In a quick test on my laptop, this makes the build about 2x faster, and shrinks the binaries from about 14MB to 9MB (after running "strip").
-
由 Adam Lee 提交于
Signed-off-by: NGang Xiong <gxiong@pivotal.io>
-
-
-
- In ORCA, due to the way exception handled previously we do warning first and then later print error referring that message. In this commit, we enhanced the exception handling so we just print a single error message. - Also, we removed 'PQO unable to generate a plan' or 'Aborting PQO plan generation' message and make the error message as close as the planner. - Updated error message with filename and line number from where the exception is raised.
-
- 31 10月, 2016 3 次提交
-
-
由 Yandong Yao 提交于
-
由 Adam Lee 提交于
Due to 31edbadf, concatenating two integers is not supported by both GPDB and latest PostgreSQL, so we need to explicitly cast the integers in this tpch_query case. Signed-off-by: NPengzhou Tang <ptang@pivotal.io> Signed-off-by: NGang Xiong <gxiong@pivotal.io>
-
由 Kenan Yao 提交于
Environment variables of QD should not be included in the startup packet when making connections to segments. Add ICG test case to guarantee the QD env variable should not effect GUCs on segments Signed-off-by: NPengzhou Tang <ptang@pivotal.io>
-
- 30 10月, 2016 3 次提交
-
-
由 Peter Eisentraut 提交于
constant format strings, so that the compiler can more easily check the formats for correctness. Conflicts: src/backend/postmaster/postmaster.c src/interfaces/ecpg/pgtypeslib/dt_common.c
-
由 Daniel Gustafsson 提交于
Commit ff3474e2 which removed the gp_partition_* catalog functions left some auxiliary code behind which is now unused. Remove the now dead functions and while in there, put back the filename in the file headercomment to match other files.
-
由 Heikki Linnakangas 提交于
The inconsistency between using tabs and spaces caused a compiler warning from GCC 6.2, with -Wmisleading-indentation.
-
- 29 10月, 2016 3 次提交
-
-
由 foyzur 提交于
* Removing invalid assertion that checks to ensure we don't do further allocations under dead accounts. The whole idea of a index based accounting is to not care about dead accounts and transparently handle those via RolloverAccount without hitting a SEGV. * Including MEMORY_OWNER_TYPE_EXECUTOR_END as a valid owner type as this enum value overlaps with the last valid executor account type.
-
These .sql files assumed that the timezone for machines running the TINC tests would have the timezone as PST/PDT. This commit explicitly sets the timezone to PST in the .sql files and updates the .ans files to handle this new behavior. This was done so that the below TINC tests could run on Concourse containers which use UTC time zone: -cs-storage -cs-uao
-
由 Nikos Armenatzoglou 提交于
Current codegen expression evaluation framework generates code based only on ExprState->Expr->type (e.g., T_Var), without checking the ExprState->type. In other words, it treats all nodes of the tree that have the same type (e.g., T_Var) in the same way (i.e., generates the same code). However, in GPDB the evaluation function (evalfunc) of a node depends on ExprState->Expr->type and ExprState->type. For example, if ExprState->Expr->type = T_Var and ExprState->type = T_ExprState, then the evalfunc = ExecEvalScalarVar. On the other hand, if ExprState->Expr->type = T_Var and ExprState->type = T_WholeRowVarExprState, then evalfunc = ExecEvalWholeRowVar. We fix this issue by avoiding code generation for an expression tree that contains nodes of unsupported ExprState->type. Signed-off-by: NKaren Huddleston <khuddleston@pivotal.io>
-
- 28 10月, 2016 13 次提交
-
-
由 Daniel Gustafsson 提交于
The tracking of process startups was in a past commit encompassed within an if brace block seemingly by accident. This seems to have been copy-pasted with the result being that tracking processes was depending on the debug_print_server_processes GUC being set. This separates the tracking and the process checking and also removes a few checks that weren't necessary.
-
由 Daniel Gustafsson 提交于
When Debug_print_server_processes is set we must send to LOG rather than DEBUG2.
-
由 Daniel Gustafsson 提交于
Log the sending of a signal regardless of if we know the name of the process or not in signal_child() when Debug_print_server_processes is set to on.
-
由 Daniel Gustafsson 提交于
Wholesale cleanup of the postmaster.c file as it was worked on. Some of the fixups include: function return value declarations on their own line; breaking long lines; concatenating broken lines that collapsed are less than 79 chars; fixing spelling; moving MPP-XXX markers to end of the comments rather then prefixing the comment; ..and more. Since the file originates from upstream, opt for the postgres style when in conflict to keep the file consistent.
-
由 Daniel Gustafsson 提交于
The returned string is better declared as a const char pointer, update and fix the two call sites. Also add the walwriter process to the list and make the if/else chain consistent by just using if. Any if statement being true will leave the function anyways so might as well make it easier to read.
-
由 Daniel Gustafsson 提交于
When constructing the version string to report in the postmaster, use safe string operators strlcpy() and strlcat(). Even though the buffer is large enough to not be likely to overflow the mere chance is worth avoiding by all means.
-
由 Daniel Gustafsson 提交于
In case an illegal option is passed, exit(1) rather than trying to keep going. Also add the default: case as well as suppress printing an error of our own since getopt() has already done so at this point.
-
由 Daniel Gustafsson 提交于
getopt() return integer and subtle differences can exist between using an integer and a char. Reported by github user liruto.
-
由 Daniel Gustafsson 提交于
If the message is truncated in construction we need to treat it as a failed construction since we otherwise might leave out important parts.
-
由 Nikos Armenatzoglou 提交于
Enhancing PGFuncGeneratorInterface, PGIRBuilderFuncGenerator, and PGGenericFuncGenerator with attributes and checks that record and examine if a built-in function is strict or not and generate code accordingly. Authors: Nikos Armenatzoglou, Nikhil Kak, Jimmy Yih, and Karen Huddleston.
-
由 Daniel Gustafsson 提交于
Commit 96f22dac expanded the feature check to handle more than just ORCA but when moving to using passed in string variables for printing on/off it generates a clang compiler warning on -Wformat-security: warning: format string is not a string literal (potentially insecure) Move to using format strings rather than just passing the string variable to fix the warning and also avoid any theoretical issues with the passed string.
-
由 Daniel Gustafsson 提交于
This is a collection of small fixes to check_feature_status() that fell out when fixing a compiler warning in the code. The fixes are (listed in no specific order): * Print tuples only in SHOW <feature>; command since we know there is little point in reading the header rows when we're just hunting for the on/off * Move to const char arguments since this function has no business altering the input * Explicitly clean up temporary files from feature checking and remove them from .gitignore files to avoid littering the tree * Check for overflow in command generation since we are appending a string with a theoretical limit of MAXPGPATH onto a string with MAXPGPATH size combined with additional content thus risking a truncation. Also fix a minor related style issue to match the rest of the file
-
由 Ashwin Agrawal 提交于
--$DEFAULT_QD_MAX_CONNECT--- Can be set as an environment variable before calling make cluster. Defalt value is 25. ---$NUM_PRIMARY_MIRROR_PAIRS--- Can be set as an environment variable before calling make cluster. gpdemo by default creates 3 primaries and 3 mirrors. With this change, users can set the environment variable $NUM_PRIMARY_MIRROR_PAIRS to another value such as 2 to make gpdemo create 2 primaries and 2 mirrors instead.
-
- 27 10月, 2016 7 次提交
-
-
由 Adam Lee 提交于
Signed-off-by: NGang Xiong <gxiong@pivotal.io>
-
由 Adam Lee 提交于
Also update related makefiles. Signed-off-by: NGang Xiong <gxiong@pivotal.io>
-
由 Ashwin Agrawal 提交于
-
由 Ashwin Agrawal 提交于
-
由 Ashwin Agrawal 提交于
Now that test code is present with the product code, stop performing version check explicitely to skip tests not relevant for specific version.
-
由 Larry Hamel 提交于
* refactor gpcheckcat foreign key validation to improve readability
-
由 Nikos Armenatzoglou 提交于
Currently we enroll ExecVariableList in InitScanStateRelationDetails. This is not the proper place to enroll it, since InitScanStateRelationDetails is called from DynamicScan_InitSingleRelation DynamicScan_Begin (i.e., Dynamic partition scan). Existing codegen framework does not support ExecVariableList for Dynamic Partition Scan. So, we move ExecVariableList's enrollment to ExecInitNode, and we enroll it only if scan is one of the next types: T_SeqScan, T_AppendOnlyScan, T_AOCSScan, or T_TableScan. Signed-off-by: NKaren Huddleston <khuddleston@pivotal.io>
-