infra-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sebb (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (INFRA-15929) Populate mboxer with historic mail archives
Date Sat, 03 Feb 2018 16:02:00 GMT

    [ https://issues.apache.org/jira/browse/INFRA-15929?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16342794#comment-16342794
] 

Sebb edited comment on INFRA-15929 at 2/3/18 4:01 PM:
------------------------------------------------------

Once INFRA-15928 is completed, it will be known from when mboxer has a complete record.
Suppose all mails from Jan 2018 are known to be present on mboxer.

All emails before then (i.e. up to Dec 2017) month can be dropped from mboxer.
[This should include all the private mails that were originally stored under /x1/archives]

The older mboxes (up to Dec 2017) can then be copied across.

Checks then need to be made to see if there are any missing or duplicated mails from the boundary
of the cut-off month:
- Missing mails will have to be recovered from Jan 2018 mbox on minotaur.
- Duplicated mails will have to be removed from one of the mboxes (Dec 2017 or Jan 2018) on
mboxer.


was (Author: sebb@apache.org):
Once INFRA-15928 is completed, it will be known from when mboxer has a complete record.
Suppose all mails from Jan 2018 are known to be present on mboxer.

All emails before then (i.e. up to Dec 2017) month can be dropped from mboxer.
[This should include all the private mails that were originally stored under /x1/archives]

The older mboxes (up to Dec 2017) can then be copied across.

Checks then need to be made to see if there are any missing or duplicated mails from the boundary
of the cut-off month.

Missing mails will have to be recovered from Jan 2018 mbox on minotaur.
Duplicated mails will have to be removed from one of the mboxes (Dec 2017/Jan 2018) on mboxer
.

> Populate mboxer with historic mail archives
> -------------------------------------------
>
>                 Key: INFRA-15929
>                 URL: https://issues.apache.org/jira/browse/INFRA-15929
>             Project: Infrastructure
>          Issue Type: Task
>          Components: Mail Archives
>            Reporter: Sebb
>            Priority: Major
>
> The mboxer VM has been set up to capture mails sent to mailing lists.
> If it is to be used as the source of all mail, it needs to include historic mailboxes
as well.
> So there needs to be a plan to copy the existing mailboxes from minotaur.
> It is unfortunately not just a case of copying the historic mbox files from minotaur
to mboxer.
> The plan needs to take into account:
> - there was a short period when mails were not being captured by minotaur and only existed
on mboxer. I think this has already been handled by back-copying the files, but this needs
to be double-checked before overwriting any files on mboxer. (See INFRA-15979)
> - the private-arch directory structure on minotaur is different from mboxer
> - the month-end changeover times may vary slightly between the two (i.e. mails sent around
the month change may end up in different files). [This is unavoidable.]
> - there were some mails dropped by mboxer in the early stages; also private mails were
mixed in with public mails. It needs to be determined at what stage the mboxer files started
to be complete.
> The mails stored on mboxer have body From_ lines properly prefixed. This is not (yet)
the case for mails on minotaur. This does not directly affect the copying strategy, but needs
to be borne in mind if comparing mail contents.
> Preconditiions:
> - ensure that all mailing lists and aliases are being correctly captured (INFRA-15928)



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

Mime
View raw message