[GitHub] incubator-trafodion pull request #758: [TRAFODION-2278] rmscheck should use ...
GitHub user mkby opened a pull request: https://github.com/apache/incubator-trafodion/pull/758 [TRAFODION-2278] rmscheck should use relative path rmscheck should use relative path of sql file location. You can merge this pull request into a Git repository by running: $ git pull https://github.com/mkby/incubator-trafodion master Alternatively you can review and apply these changes as the patch at: https://github.com/apache/incubator-trafodion/pull/758.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 #758 commit 56e498ff081b70d967f83f5379c197ef2d3db3b1 Author: EasonDate: 2016-10-12T03:11:41Z [TRAFODION-2278] rmscheck should use relative path rmscheck should use relative path of sql file location. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #757: [TRAFODION-2167]: Invalid query inval...
GitHub user robertamarton opened a pull request: https://github.com/apache/incubator-trafodion/pull/757 [TRAFODION-2167]: Invalid query invalidation keys not working properly When a user is revoked from a role, invalidation keys are not being processed correctly. Therefore, users can still run queries even though privileges have been removed. Query invalidation is complicated when table descriptors are stored in metadata. Changes: --> The list of priv_descs created (and stored) was changed to include an entry for each user and each role accumulated across all grantors. (Today, each priv_desc entry includes the users' direct grants plus grants on their active roles.) --> When an object is loaded into NATable or NARoutine cache, the priv_desc is accessed and the privilege bitmap is now generated by combining the users' privileges with privileges of their active roles. Correct invalidation keys are now being created and stored with the object. In the first code drop, the users' active roles are read from the role_usage table. In the next code drop, the active roles will be stored and maintained in executor globals. --> When a plan is compiled, the correct invalidation keys for users, roles, and the public authorization are added to the plan. --> Changes in the compiler were required to handle the invalidation keys for revoke role and revoke privilege from "PUBLIC". --> Cleaned up the code that manages invalidation keys in privilege manager. --> Included the correct create and redef times (if available) in the stored object descriptor - today they are always set to 0. --> Added new regression test to test all the revoke options. You can merge this pull request into a Git repository by running: $ git pull https://github.com/robertamarton/incubator-trafodion trafodion-2189 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/incubator-trafodion/pull/757.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 #757 commit 3b437720a54de88b6a04e44e0257fac0448f0762 Author: Roberta MartonDate: 2016-10-12T01:48:22Z [TRAFODION-2167]: Invalid query invalidation keys not working properly When a user is revoked from a role, invalidation keys are not being processed correctly. Therefore, users can still run queries even though privileges have been removed. Query invalidation is complicated when table descriptors are stored in metadata. Changes: --> The list of priv_descs created (and stored) was changed to include an entry for each user and each role accumulated across all grantors. (Today, each priv_desc entry includes the users' direct grants plus grants on their active roles.) --> When an object is loaded into NATable or NARoutine cache, the priv_desc is accessed and the privilege bitmap is now generated by combining the users' privileges with privileges of their active roles. Correct invalidation keys are now being created and stored with the object. In the first code drop, the users' active roles are read from the role_usage table. In the next code drop, the active roles will be stored and maintained in executor globals. --> When a plan is compiled, the correct invalidation keys for users, roles, and the public authorization are added to the plan. --> Changes in the compiler were required to handle the invalidation keys for revoke role and revoke privilege from "PUBLIC". --> Cleaned up the code that manages invalidation keys in privilege manager. --> Included the correct create and redef times (if available) in the stored object descriptor - today they are always set to 0. --> Added new regression test to test all the revoke options. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #743: [TRAFODION-2259] TopN sort changes.
Github user asfgit closed the pull request at: https://github.com/apache/incubator-trafodion/pull/743 --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #756: Trafodion 2235
GitHub user zcorrea opened a pull request: https://github.com/apache/incubator-trafodion/pull/756 Trafodion 2235 Fix to Zookeeper client logic to ignore ZNODEEXISTS errors when creating directory tree structure. You can merge this pull request into a Git repository by running: $ git pull https://github.com/zcorrea/incubator-trafodion TRAFODION-2235 Alternatively you can review and apply these changes as the patch at: https://github.com/apache/incubator-trafodion/pull/756.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 #756 commit 6dc469df784583ac3bd8df28d61386d73517cf10 Author: Zalo CorreaDate: 2016-09-27T16:14:24Z Merge remote branch 'origin/master' into TRAFODION-2235 commit 41c384ed5413c91c80b368bf0dfb74daef7ca1ab Author: Zalo Correa Date: 2016-10-11T17:31:37Z Merge remote branch 'origin/master' into TRAFODION-2235 commit 3d65adcb3fcbc5c92f6ca18f413bd374b36827a0 Author: Zalo Correa Date: 2016-10-11T19:57:06Z TRAFODION-2235 - Fixed zclient logic to ignore ZNODEEXISTS error when creating directory tree. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #:
Github user prashanth-vasudev commented on the pull request: https://github.com/apache/incubator-trafodion/commit/3cd201a6eb98ec412e33af44f106b69b215e2dcf#commitcomment-19381216 In core/sql/sort/Topn.cpp: In core/sql/sort/Topn.cpp on line 121: This has been fixed in the next commit --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #:
Github user selvaganesang commented on the pull request: https://github.com/apache/incubator-trafodion/commit/3cd201a6eb98ec412e33af44f106b69b215e2dcf#commitcomment-19381197 In core/sql/sort/Topn.cpp: In core/sql/sort/Topn.cpp on line 121: It looks like Record is in the general C heap but the contents could be in heap_ passed in. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #754: [TRAFODION-2262] Mxosrvr or java core...
Github user DaveBirdsall commented on a diff in the pull request: https://github.com/apache/incubator-trafodion/pull/754#discussion_r8280 --- Diff: core/sql/qmscommon/QRLogger.cpp --- @@ -186,6 +188,9 @@ NABoolean QRLogger::initLog4cxx(const char* configFileName) { NAString logFileName; + if (gv_QRLoggerInitialized_) + return TRUE; + // get the log directory --- End diff -- I cannot figure out if this code is thread-safe. What if two threads concurrently enter this function and both see gv_QRLoggerInitialized_ as FALSE. Both would then call CommonLogger::initLog4cxx and possibly introduceSelf(). Is there any harm in this? --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---
[GitHub] incubator-trafodion pull request #753: [TRAFODION-2272] Insure adequate hbas...
Github user selvaganesang commented on a diff in the pull request: https://github.com/apache/incubator-trafodion/pull/753#discussion_r82815496 --- Diff: core/sql/src/main/java/org/trafodion/sql/HBaseClient.java --- @@ -159,6 +159,17 @@ public HBaseClient() { static public Connection getConnection() throws IOException { +// On some distributions, the hbase.client.scanner.timeout.period setting is +// too small, resulting in annoying SocketTimeoutExceptions during operations +// such as UPDATE STATISTICS on very large tables. On CDH 5.4.5 in particular +// we have seen this. Unfortunately Cloudera Manager does not allow us to +// change this setting, and setting it manually in hbase-site.xml doesn't work +// because a later Cloudera Manager deploy would just overwrite it. So, we +// programmatically check the setting here and insure it is at least 1 hour. +long configuredTimeout = config.getLong("hbase.client.scanner.timeout.period",0); +if (configuredTimeout < 360 /* 1 hour */) + config.setLong("hbase.client.scanner.timeout.period",360); --- End diff -- Connection is also created at line 184 also. Can you please change it to call this method so that all connections have this property configured. --- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. ---