[
https://issues.apache.org/jira/browse/ZOOKEEPER-882?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12972667#action_12972667
]
Jared Cantwell commented on ZOOKEEPER-882:
------------------------------------------
Flavio, so I figured out what was holding me up before. In an earlier comment
you said the following, "Regarding init(), we first use the value of zxid to
determine which log files to read: all log files tagged with a value higher
than zxid and the last log file that is less than zxid. Next we iterate over
the log files until hdr.getZxid() is greater or equal to zxid (should be zxid
really). This guarantees that the next call to next(), after init() returns,
will return zxid+1." This is true almost all the time. It is true when we
have the requested zxid in our log, but NOT if the very first log entry is
zxid+1, which can easily be the case in almost any startup scenario. I am
betting that the loop in restore() was originally planned to be coded as you
stated (by initializing with last lastProcessedTxnID and then calling next()),
but then the issue I mentioned came up and it was decided it was ok to just use
the first header, and call next afterward. Without rewriting parts of the
TxnIterator, I don't think restructuring that loop in restore() is simple.
Therefore, I propose that we simply request lastProcessedZxid+1 from the log
initially (which is guaranteed to either be present, or have no entries at all
in the log). I am about to submit a patch, with a unittest for this.
> Startup loads last transaction from snapshot
> --------------------------------------------
>
> Key: ZOOKEEPER-882
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-882
> Project: ZooKeeper
> Issue Type: Bug
> Components: server
> Reporter: Jared Cantwell
> Assignee: Jared Cantwell
> Priority: Minor
> Fix For: 3.4.0
>
> Attachments: 882.diff, FailureTest-882.patch, restore,
> ZOOKEEPER-882.patch, ZOOKEEPER-882.patch
>
>
> On startup, the server first loads the latest snapshot, and then loads from
> the log starting at the last transaction in the snapshot. It should begin
> from one past that last transaction in the log. I will attach a possible
> patch.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
