[
https://issues.apache.org/jira/browse/HDFS-3049?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13278341#comment-13278341
]
Colin Patrick McCabe commented on HDFS-3049:
--------------------------------------------
bq. Also, I'm not sure why the exception is swallowed instead of rethrown. If
it fails to open the edit log, shouldn't that generate an exception on init()?
Should we make init() a public interface (eg "open()") instead, so that the
caller is cognizant of the flow here, instead of doing it lazily? I think that
would also simplify the other functions, which could just do a
Preconditions.checkState(state == State.OPEN) instead of always handling
lazy-init.
You're right-- the exception should be thrown if resync == false.
As for creating a public init() method-- I guess, but this patch is getting
kind of big already. Perhaps we could file a separate JIRA for that? I also
have some other API ideas that might improve efficiency (not going to discuss
them here due to space constraints)
bq. Why do you close() here but not close() in the normal case where you reach
the end of the log? It seems it should be up to the caller to close upon
hitting the "eof" (null txn) either way.
The rationale behind this was discussed in HDFS-3335. Basically, if there is
corruption at the end of the log, but we read everything we were supposed to,
we don't want to throw an exception. As for closing in the eof case, that
seems unecessary. The caller has to call close() anyway, that's part of the
contract for this API. So we don't really add any value by doing it
automatically.
bq. again, why not just catch Throwable?
Yeah, we should do that. Will fix.
bq. IllegalStateException would be more appropriate here
ok
[streamlimiter comments]
agree with most/all of this. I think this can be separated out (probably)
[log message comments]
yes, probably some of those should be debug comments. Probably at least the
ones which just describe "situation normal, added new stream" etc.
[separate into 3 patches]
well, it's worth a try. There are some non-obvious dependencies, but I'll give
it a try.
> During the normal loading NN startup process, fall back on a different
> EditLog if we see one that is corrupt
> ------------------------------------------------------------------------------------------------------------
>
> Key: HDFS-3049
> URL: https://issues.apache.org/jira/browse/HDFS-3049
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: name-node
> Affects Versions: 0.23.0
> Reporter: Colin Patrick McCabe
> Assignee: Colin Patrick McCabe
> Priority: Minor
> Attachments: HDFS-3049.001.patch, HDFS-3049.002.patch,
> HDFS-3049.003.patch, HDFS-3049.005.against3335.patch,
> HDFS-3049.006.against3335.patch, HDFS-3049.007.against3335.patch,
> HDFS-3049.010.patch, HDFS-3049.011.patch, HDFS-3049.012.patch,
> HDFS-3049.013.patch, HDFS-3049.015.patch, HDFS-3049.017.patch,
> HDFS-3049.018.patch
>
>
> During the NameNode startup process, we load an image, and then apply edit
> logs to it until we believe that we have all the latest changes.
> Unfortunately, if there is an I/O error while reading any of these files, in
> most cases, we simply abort the startup process. We should try harder to
> locate a readable edit log and/or image file.
> *There are three main use cases for this feature:*
> 1. If the operating system does not honor fsync (usually due to a
> misconfiguration), a file may end up in an inconsistent state.
> 2. In certain older releases where we did not use fallocate() or similar to
> pre-reserve blocks, a disk full condition may cause a truncated log in one
> edit directory.
> 3. There may be a bug in HDFS which results in some of the data directories
> receiving corrupt data, but not all. This is the least likely use case.
> *Proposed changes to normal NN startup*
> * We should try a different FSImage if we can't load the first one we try.
> * We should examine other FSEditLogs if we can't load the first one(s) we try.
> * We should fail if we can't find EditLogs that would bring us up to what we
> believe is the latest transaction ID.
> Proposed changes to recovery mode NN startup:
> we should list out all the available storage directories and allow the
> operator to select which one he wants to use.
> Something like this:
> {code}
> Multiple storage directories found.
> 1. /foo/bar
> edits__curent__XYZ size:213421345 md5:2345345
> image size:213421345 md5:2345345
> 2. /foo/baz
> edits__curent__XYZ size:213421345 md5:2345345345
> image size:213421345 md5:2345345
> Which one would you like to use? (1/2)
> {code}
> As usual in recovery mode, we want to be flexible about error handling. In
> this case, this means that we should NOT fail if we can't find EditLogs that
> would bring us up to what we believe is the latest transaction ID.
> *Not addressed by this feature*
> This feature will not address the case where an attempt to access the
> NameNode name directory or directories hangs because of an I/O error. This
> may happen, for example, when trying to load an image from a hard-mounted NFS
> directory, when the NFS server has gone away. Just as now, the operator will
> have to notice this problem and take steps to correct it.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
