Last modified: 2014-10-29 17:09:43 UTC
The bits partition [1] for 2014-10-20T02/1H has not been marked successful. What happened? [1] _________________________________________________________________ qchris@stat1002 // jobs: 0 // time: 09:04:05 // exit code: 0 cwd: ~ ~/cluster-scripts/dump_webrequest_status.sh +------------------+--------+--------+--------+--------+ | Date | bits | mobile | text | upload | +------------------+--------+--------+--------+--------+ [...] | 2014-10-20T08/1H | . | . | . | . | | 2014-10-20T09/1H | . | . | . | . | | 2014-10-20T10/1H | X | . | . | . | | 2014-10-20T11/1H | . | . | . | . | | 2014-10-20T12/1H | . | . | . | . | [...] +------------------+--------+--------+--------+--------+ Statuses: . --> Partition is ok M --> Partition manually marked ok X --> Partition is not ok (duplicates, missing, or nulls)
The affected period is 10:37:16--10:37:20, which nicely matches the manual kafka leader re-election from 10:38. Mismatching data is minimal: +----------------------------+-----------+--------------+ | Host | # missing | # duplicates | +----------------------------+-----------+--------------+ | cp3019.esams.wikimedia.org | 0 | 252 | | cp3020.esams.wikimedia.org | 0 | 252 | | cp3021.esams.wikimedia.org | 525 | 0 | | cp4004.ulsfo.wmnet | 220 | 0 | +----------------------------+-----------+--------------+ Total worth of mismatched data <<1 second.
