[ https://issues.apache.org/jira/browse/RANGER-3766?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Abhinaba Sarkar updated RANGER-3766: ------------------------------------ Attachment: createrangerdb4.log > Connection issues during execution of db_setup.py wipes shared DB > ----------------------------------------------------------------- > > Key: RANGER-3766 > URL: https://issues.apache.org/jira/browse/RANGER-3766 > Project: Ranger > Issue Type: Bug > Components: admin > Affects Versions: 1.2.0 > Reporter: Abhinaba Sarkar > Priority: Major > Attachments: createrangerdb4.log > > > While running db_setup.py on a shared ranger DB (schema already present), if > there is an intermittent connection failure while executing the following > lines (351-354)- > {code:java} > isFirstTableExist = self.check_table(db_name, db_user, db_password, > first_table) isLastTableExist = > self.check_table(db_name, db_user, db_password, last_table) > > isDBPatchesApplied=self.hasDBnJavaPatchesEntries(db_name, db_user, > db_password, "DB_PATCHES") > isJavaPatchesApplied=self.hasDBnJavaPatchesEntries(db_name, db_user, > db_password, "JAVA_PATCHES") {code} > the entire schema is re-imported and all the previous data is lost. > This happens because while executing the command through > {code:java} > check_output(){code} > method, we ignore the stderr. In case of any error since stdout is empty, the > code assumes that there is no existing schema and re-imports the schema file. > Issue found on version 1.2. Tested with MSSQL server on azure hdinsight > > This issue can be reproduced by debugging using pdb and blocking connection > to the DB when executing the above lines. Logs attached. -- This message was sent by Atlassian Jira (v8.20.7#820007)