[ https://issues.apache.org/jira/browse/RATIS-1887 ]
GuoHao deleted comment on RATIS-1887:
-------------------------------
was (Author: nicholas niu):
Is it possible to solve this problem by deleting the `ToDelete` segemnt log
first, then trancate & rename the `ToTruncate` segment log.
Like this:
{code:java}
//
org.apache.ratis.server.raftlog.segmented.SegmentedRaftLogWorker.TruncateLog#execute
@Override
void execute() throws IOException {
freeSegmentedRaftLogOutputStream();
if (segments.getToDelete() != null && segments.getToDelete().length > 0) {
long minStart = segments.getToDelete()[0].getStartIndex();
for (SegmentFileInfo del : segments.getToDelete()) {
final File delFile = del.getFile(storage);
Preconditions.assertTrue(delFile.exists(),
"File %s to be deleted does not exist", delFile);
FileUtils.deleteFile(delFile);
LOG.info("{}: Deleted log file {}", name, delFile);
minStart = Math.min(minStart, del.getStartIndex());
}
if (segments.getToTruncate() == null) {
lastWrittenIndex = minStart - 1;
}
}
if (segments.getToTruncate() != null) {
final File fileToTruncate = segments.getToTruncate().getFile(storage);
Preconditions.assertTrue(fileToTruncate.exists(),
"File %s to be truncated does not exist", fileToTruncate);
FileUtils.truncateFile(fileToTruncate,
segments.getToTruncate().getTargetLength());
// rename the file
final File dstFile = segments.getToTruncate().getNewFile(storage);
Preconditions.assertTrue(!dstFile.exists(),
"Truncated file %s already exists ", dstFile);
FileUtils.move(fileToTruncate, dstFile);
LOG.info("{}: Truncated log file {} to length {} and moved it to {}", name,
fileToTruncate, segments.getToTruncate().getTargetLength(), dstFile);
// update lastWrittenIndex
lastWrittenIndex = segments.getToTruncate().getNewEndIndex();
}
if (stateMachineFuture != null) {
IOUtils.getFromFuture(stateMachineFuture, () -> this +
"-truncateStateMachineData");
}
flushIndex.setUnconditionally(lastWrittenIndex, infoIndexChange);
safeCacheEvictIndex.setUnconditionally(lastWrittenIndex, infoIndexChange);
postUpdateFlushedIndex(0);
}{code}
> Gap between segement log
> ------------------------
>
> Key: RATIS-1887
> URL: https://issues.apache.org/jira/browse/RATIS-1887
> Project: Ratis
> Issue Type: Bug
> Reporter: GuoHao
> Priority: Critical
> Attachments: image-2023-09-08-10-18-36-198.png
>
>
>
> My version of ratis already includes this
> pr(https://issues.apache.org/jira/browse/RATIS-1763) and I am using a new
> raft server.
>
> Description:
> 1. i am using ratis version 2.5.1
> 2. the application software is ozone 1.3.0 scm ha
>
> scm error log:
> {code:java}
> Caused by: java.lang.IllegalStateException: Found a gap between logs: the
> last log segment log-784848_809981 ended at 809981 but the next log segment
> log-822560_856038 started at 822560
> at
> org.apache.ratis.util.Preconditions.assertTrue(Preconditions.java:72)
> at
> org.apache.ratis.server.raftlog.segmented.SegmentedRaftLogCache.validateAdding(SegmentedRaftLogCache.java:424)
> at
> org.apache.ratis.server.raftlog.segmented.SegmentedRaftLogCache.addSegment(SegmentedRaftLogCache.java:431)
> at
> org.apache.ratis.server.raftlog.segmented.SegmentedRaftLogCache.loadSegment(SegmentedRaftLogCache.java:384)
> at
> org.apache.ratis.server.raftlog.segmented.SegmentedRaftLog.loadLogSegments(SegmentedRaftLog.java:241)
> at
> org.apache.ratis.server.raftlog.segmented.SegmentedRaftLog.openImpl(SegmentedRaftLog.java:214)
> at
> org.apache.ratis.server.raftlog.RaftLogBase.open(RaftLogBase.java:251)
> at
> org.apache.ratis.server.impl.ServerState.initRaftLog(ServerState.java:239)
> at
> org.apache.ratis.server.impl.ServerState.initRaftLog(ServerState.java:220)
> at
> org.apache.ratis.server.impl.ServerState.lambda$new$5(ServerState.java:161)
> at
> org.apache.ratis.util.MemoizedSupplier.get(MemoizedSupplier.java:62)
> at
> org.apache.ratis.server.impl.ServerState.initialize(ServerState.java:177)
> at
> org.apache.ratis.server.impl.RaftServerImpl.start(RaftServerImpl.java:338)
> at
> org.apache.ratis.util.ConcurrentUtils.accept(ConcurrentUtils.java:188){code}
> segment log:
>
> !image-2023-09-08-10-18-36-198.png!
>
> The modification time of this segment log is greater than the modification
> time of the file with the larger index.
> The file size of this file seems to be larger than the other files, but it's
> not as big as the other files.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
