[ https://issues.apache.org/jira/browse/HBASE-16235?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ted Yu updated HBASE-16235: --------------------------- Hadoop Flags: Reviewed Fix Version/s: 2.0.0 > TestSnapshotFromMaster#testSnapshotHFileArchiving will fail if there are too > many hfiles > ---------------------------------------------------------------------------------------- > > Key: HBASE-16235 > URL: https://issues.apache.org/jira/browse/HBASE-16235 > Project: HBase > Issue Type: Bug > Reporter: ChiaPing Tsai > Assignee: ChiaPing Tsai > Priority: Trivial > Fix For: 2.0.0 > > Attachments: HBASE-16235-v1.patch, HBASE-16235-v2.patch, > HBASE-16235-v3.patch > > > TestSnapshotFromMaster#testSnapshotHFileArchiving assumes that all hfiles > will be compacted and be moved to “archive folder” after cleaning. But not > all hfiles will be compacted if there are large number of hfiles. > The above may be happened if changing the default config like smaller write > buffer(hbase.client.write.buffer) or ExponentialClientBackoffPolicy. > {code:title=TestSnapshotFromMaster.java|borderStyle=solid} > // it should also check the hfiles in the normal path > (/hbase/data/default/...) > public void testSnapshotHFileArchiving() throws Exception { > //... > // get the archived files for the table > Collection<String> files = getArchivedHFiles(archiveDir, rootDir, fs, > TABLE_NAME); > // and make sure that there is a proper subset > for (String fileName : snapshotHFiles) { > assertTrue("Archived hfiles " + files + " is missing snapshot file:" + > fileName, > files.contains(fileName)); > } > //... > } > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)