[jira] [Comment Edited] (OAK-9256) Missing segment not detected in AzureSegmentArchiveReader

["Miroslav Smiljanic (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/OAK-9256?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17220586#comment-17220586 ] 

Miroslav Smiljanic edited comment on OAK-9256 at 10/26/20, 10:36 AM:
---------------------------------------------------------------------

Hi [~adulceanu]

 
{quote} * in {{AzureUtilities.readBufferFully()}}, in order to keep the code simple, we could just check the http status code coming directly from the exception with {{e.getHttpStatusCode}}(){quote}
Good suggestion (y), haven't seen before that the same info is available in the exception. 

Modified patch: [^OAK-9256_2.patch]
{quote} * It's not clear to me why we need to introduce the new {{FileNotFoundInArchiveException}}, when we already have the {{SegmentNotFoundException}}. I would just re-use the latter and change accordingly the test case(s).{quote}
Method org.apache.jackrabbit.oak.segment.azure.AzureUtilities#readBufferFully is also used in [AzureSegmentArchiveReader#doReadDataFile|https://github.com/apache/jackrabbit-oak/blob/jackrabbit-oak-1.34.0/oak-segment-azure/src/main/java/org/apache/jackrabbit/oak/segment/azure/AzureSegmentArchiveReader.java#L80] and JavaDoc says:
{noformat}
/**
 * Reads a data file inside the archive. This entry is not a segment. Its full name is given by archive name + extension.
{noformat}
This is the reason why I thought that it would make sense to introduce FileNotFoundInArchiveException, as a super class for SegmentNotFoundException.


was (Author: smiroslav):
Hi [~adulceanu]

 
{quote} * in {{AzureUtilities.readBufferFully()}}, in order to keep the code simple, we could just check the http status code coming directly from the exception with {{e.getHttpStatusCode}}(){quote}
Good suggestion (y), haven't seen before that the same info is available in the exception. 

Modified patch: [^OAK-9256_2.patch]
{quote} * It's not clear to me why we need to introduce the new {{FileNotFoundInArchiveException}}, when we already have the {{SegmentNotFoundException}}. I would just re-use the latter and change accordingly the test case(s).{quote}
Method org.apache.jackrabbit.oak.segment.azure.AzureUtilities#readBufferFully is also used in [AzureSegmentArchiveReader#doReadDataFile|https://github.com/apache/jackrabbit-oak/blob/jackrabbit-oak-1.34.0/oak-segment-azure/src/main/java/org/apache/jackrabbit/oak/segment/azure/AzureSegmentArchiveReader.java#L80] and JavaDoc says:
{noformat}
/**
 * Reads a data file inside the archive. This entry is not a segment. Its full name is given by archive name + extension.
{noformat}
This is the reason why I thought that it would make sens to introduce FileNotFoundInArchiveException, as a super class for SegmentNotFoundException.

> Missing segment not detected in AzureSegmentArchiveReader
> ---------------------------------------------------------
>
>                 Key: OAK-9256
>                 URL: https://issues.apache.org/jira/browse/OAK-9256
>             Project: Jackrabbit Oak
>          Issue Type: Bug
>          Components: segment-azure
>    Affects Versions: 1.24.0, 1.34.0
>            Reporter: Miroslav Smiljanic
>            Assignee: Andrei Dulceanu
>            Priority: Major
>         Attachments: OAK-9256.patch, OAK-9256_2.patch, test_case.patch
>
>
> If segment has been deleted after segment reader has been initiated, that can go unnoticed and instead of reporting missing segment, exception RepositoryNotReachableException is thrown, which does not correctly describe the situation.
> Test case attached: [^test_case.patch]
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)