[ https://issues.apache.org/jira/browse/CASSANDRA-7066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14700533#comment-14700533 ]
Stefania commented on CASSANDRA-7066: ------------------------------------- I agree on stashing files with {{ignore}}, not sure for the case when we refuse to start though. Perhaps in this case we should add this functionality to the offline sstable utility tool and let the operator either clean-up or stash. I've opened CASSANDRA-10112, let's move the discussion there. [~benedict] could you commit the fix for two COVERITY defects? Commit [here|https://github.com/stef1927/cassandra/commit/c94d8a12dc43cd4705f1aa8cf384fb8c3290a5f9]. {code} ** CID 1316515: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) /src/java/org/apache/cassandra/db/lifecycle/TransactionLog.java: 363 in org.apache.cassandra.db.lifecycle.TransactionLog$TransactionFile.readRecord(java.lang.String, boolean)() ________________________________________________________________________________________________________ *** CID 1316515: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) /src/java/org/apache/cassandra/db/lifecycle/TransactionLog.java: 363 in org.apache.cassandra.db.lifecycle.TransactionLog$TransactionFile.readRecord(java.lang.String, boolean)() 357 if (!matcher.matches() || matcher.groupCount() != 2) 358 { 359 handleReadRecordError(String.format("cannot parse line \"%s\"", line), isLast); 360 return Record.make(line, isLast); 361 } 362 >>> CID 1316515: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) >>> Found reliance on default encoding: String.getBytes(). 363 byte[] bytes = matcher.group(1).getBytes(); 364 checksum.update(bytes, 0, bytes.length); 365 366 if (checksum.getValue() != Long.valueOf(matcher.group(2))) 367 handleReadRecordError(String.format("invalid line checksum %s for \"%s\"", matcher.group(2), line), isLast); 368 ** CID 1316514: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) /src/java/org/apache/cassandra/db/lifecycle/TransactionLog.java: 231 in org.apache.cassandra.db.lifecycle.TransactionLog$Record.getBytes()() ________________________________________________________________________________________________________ *** CID 1316514: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) /src/java/org/apache/cassandra/db/lifecycle/TransactionLog.java: 231 in org.apache.cassandra.db.lifecycle.TransactionLog$Record.getBytes()() 225 { 226 return String.format("%s:[%s,%d,%d]", type.toString(), relativeFilePath, updateTime, numFiles); 227 } 228 229 public byte[] getBytes() 230 { >>> CID 1316514: FindBugs: Internationalization (FB.DM_DEFAULT_ENCODING) >>> Found reliance on default encoding: String.getBytes(). 231 return record.getBytes(); 232 } 233 234 public boolean verify(String parentFolder, boolean lastRecordIsCorrupt) 235 { 236 if (type != RecordType.REMOVE) {code} > Simplify (and unify) cleanup of compaction leftovers > ---------------------------------------------------- > > Key: CASSANDRA-7066 > URL: https://issues.apache.org/jira/browse/CASSANDRA-7066 > Project: Cassandra > Issue Type: Improvement > Components: Core > Reporter: Benedict > Assignee: Stefania > Priority: Minor > Labels: benedict-to-commit, compaction > Fix For: 3.0 alpha 1 > > Attachments: 7066.txt > > > Currently we manage a list of in-progress compactions in a system table, > which we use to cleanup incomplete compactions when we're done. The problem > with this is that 1) it's a bit clunky (and leaves us in positions where we > can unnecessarily cleanup completed files, or conversely not cleanup files > that have been superceded); and 2) it's only used for a regular compaction - > no other compaction types are guarded in the same way, so can result in > duplication if we fail before deleting the replacements. > I'd like to see each sstable store in its metadata its direct ancestors, and > on startup we simply delete any sstables that occur in the union of all > ancestor sets. This way as soon as we finish writing we're capable of > cleaning up any leftovers, so we never get duplication. It's also much easier > to reason about. -- This message was sent by Atlassian JIRA (v6.3.4#6332)