[jira] [Commented] (SOLR-9536) OldBackupDirectory timestamp init bug causes NPEs from SnapShooter?

["ASF subversion and git services (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-9536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15605805#comment-15605805 ] 

ASF subversion and git services commented on SOLR-9536:
-------------------------------------------------------

Commit 22117ddde47bc5b646ec1f0e732b51479a8e8bab in lucene-solr's branch refs/heads/branch_6x from markrmiller
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=22117dd ]

SOLR-9536: OldBackupDirectory timestamp field needs to be initialized to avoid NPE.

This closes #81.


> OldBackupDirectory timestamp init bug causes NPEs from SnapShooter?
> -------------------------------------------------------------------
>
>                 Key: SOLR-9536
>                 URL: https://issues.apache.org/jira/browse/SOLR-9536
>             Project: Solr
>          Issue Type: Bug
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Hoss Man
>            Assignee: Varun Thacker
>
> On IRC, a 6.2.0 user reported getting an NPE from SnapShooter.deleteOldBackups L244, with the only other frame of the stacktrace being {{lambda$createSnapAsync$1}} L196 (it was a screenshot, not text easily cut/paste here)
> The problematic L244 is...
> {code}
>   if (obd.getTimestamp().isPresent()) {
> {code}
> ..and i believe the root of the issue is that while {{getTimestamp()}} is declared to return an {{Optional<Date>}}, there is no guarantee that the {{Optional}} instance is itself non-null...
> {code}
>    private Optional<Date> timestamp;
>   public OldBackupDirectory(URI basePath, String dirName) {
>     this.dirName = Preconditions.checkNotNull(dirName);
>     this.basePath = Preconditions.checkNotNull(basePath);
>     Matcher m = dirNamePattern.matcher(dirName);
>     if (m.find()) {
>       try {
>         this.timestamp = Optional.of(new SimpleDateFormat(SnapShooter.DATE_FMT, Locale.ROOT).parse(m.group(1)));
>       } catch (ParseException e) {
>         this.timestamp = Optional.empty();
>       }
>     }
>   }
> {code}
> Allthough i'm not 100% certain, I believe the way the user was triggering this bug was by configuring classic replication configured with something like {{<str name="replicateAfter">commit</str>}} -- so that usage may be neccessary to trigger the exception?
> Alternatively: perhaps this exception gets logged the *first* time anyone tries to use any code that involves SnapShooter -- and after that a timestamp file *is* created and teh problem neer manifests itself again?



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org