infra-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sebb (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (INFRA-16282) [mail-private] recovery needed from gzcat error
Date Mon, 02 Apr 2018 22:17:01 GMT

    [ https://issues.apache.org/jira/browse/INFRA-16282?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16423208#comment-16423208
] 

Sebb commented on INFRA-16282:
------------------------------

Note: this problem was not noticed before because the files were copied from thor where the
gz files had already been expanded. It's only now that a new month has started that the problem
has occurred.

The mail message only showed errors for the following directories:
Error: No mbox Files found in '/x1/mail-private.apache.org/mod_mbox/dubbo-security/'
Error: No mbox Files found in '/x1/mail-private.apache.org/mod_mbox/druid-private/'

This is because they only have recent files which are now all empty.

There were lots of errors of the form:
 sh: 1: gzcat: not found
but without any context they are useless (see INFRA-16281)

There must be lots of other empty 201803 files; I'll check.

> [mail-private] recovery needed from gzcat error
> -----------------------------------------------
>
>                 Key: INFRA-16282
>                 URL: https://issues.apache.org/jira/browse/INFRA-16282
>             Project: Infrastructure
>          Issue Type: Bug
>          Components: Mail Archives
>            Reporter: Sebb
>            Assignee: Sebb
>            Priority: Major
>
> The refresh-index.pl script failed to expand gz files because gzcat does not exist on
Ubuntu.
> The result was that an empty file was created.
> $ ls -l /x1/mail-private.apache.org/mod_mbox/druid-private                
> total 76
> -rwxrwxr-x 1 modmbox modmbox     0 Apr  1 00:33 201803
> -rwxrwxr-x 1 modmbox modmbox 40142 Mar 30 21:07 201803.gz
> -rwxrwxr-x 1 modmbox modmbox 24576 Mar 30 21:20 201803.mbox.msgsum
> -rwxrwxr-x 1 modmbox modmbox     0 Apr  1 00:03 201804
> -rwxrwxr-x 1 modmbox modmbox 12288 Mar 30 21:20 listinfo.db
> $ ls -l /x1/mail-private.apache.org/mod_mbox/dubbo-security/
> total 140
> -rwxrwxr-x 1 modmbox modmbox      0 Apr  1 00:33 201803
> -rwxrwxr-x 1 modmbox modmbox 115127 Mar 30 07:40 201803.gz
> -rwxrwxr-x 1 modmbox modmbox  12288 Mar 14 05:20 201803.mbox.msgsum
> -rwxrwxr-x 1 modmbox modmbox      0 Apr  1 00:03 201804
> -rwxrwxr-x 1 modmbox modmbox  12288 Mar 14 05:20 listinfo.db
> Note that the files 201803 are empty, and there is no 201803.mbox link.
> Since the files are only expanded if they are in the rsync log, and won't generally be
updated again, the expansion/indexing won't happen without intervention.
> I think the simplest way to ensure the gz file is re-expanded and indexed is to delete
the gz file. It will be fetched on the next run and processed as normal.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message