infra-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sebb (JIRA)" <j...@apache.org>
Subject [jira] [Created] (INFRA-17754) Corrupted mbox files on minotaur on Oct 25 2018
Date Thu, 24 Jan 2019 12:57:00 GMT
Sebb created INFRA-17754:
----------------------------

             Summary: Corrupted mbox files on minotaur on Oct 25 2018
                 Key: INFRA-17754
                 URL: https://issues.apache.org/jira/browse/INFRA-17754
             Project: Infrastructure
          Issue Type: Bug
          Components: Mail Archives
            Reporter: Sebb


There are several instances of corrupted mailboxes on minotaur/mailarchive

These all occured around 11:45 to 12:30 on Oct 25 2018.
They all involve a duplicated mail where the second copy is truncated.
It does not have the two CRLFs at the end.
As a result, the From_ header line for the next message does not start in column 1 and is
not therefore recognised.

The details are:

bookkeeper-commits: First: 2343 Last: 3569 Missing: [3377] Duplicates: [3376]]
bookkeeper-issues: First: 4665 Last: 14818 Missing: [13224] Duplicates: [13223]]
flink-issues: First: 145655 Last: 208989 Missing: [197434] Duplicates: [197430]]
flink-user: First: 17329 Last: 25090 Missing: [23926] Duplicates: [23925]]
mynewt-commits: First: 15229 Last: 23994 Missing: [22648] Duplicates: [22649]]
nifi-users: First: 7710 Last: 10444 Missing: [9983] Duplicates: [9982]]

In all cases the duplicates appear to be identical messages - the only difference is the data
stamp on the From_ line. The first Received line in each pair has the same date and SMTP Id.
This suggests that the duplication occured on minotaur.

It looks as if the mbox files can be fixed by removing the truncated duplicate.

But why did the issue occur?



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

Mime
View raw message