diff options
| author | Linus Nordberg <linus@nordu.net> | 2016-11-23 17:09:48 +0100 |
|---|---|---|
| committer | Linus Nordberg <linus@nordu.net> | 2016-11-23 17:09:48 +0100 |
| commit | 19a2a611a839c0318f58347e2d93943c8e2401a5 (patch) | |
| tree | 18cd302161a88d4546b39792a4bff6b1ade95c77 /doc/merge.txt | |
| parent | 27e368196ce65e109c027987c706a697356f7bc5 (diff) | |
WIP
Merge can run as four separate processes, plus a fifth controlling
proces 'merge'.
Tests are limited to testcase1.py and they're failing because of the
test with the dead merge secondary. Tests are also time consuming
because they're waiting for 60s each time a merge needs to be
verified. This could be improved by peeking at the control files, for
example.
Diffstat (limited to 'doc/merge.txt')
| -rw-r--r-- | doc/merge.txt | 9 |
1 files changed, 7 insertions, 2 deletions
diff --git a/doc/merge.txt b/doc/merge.txt index d12424f..c7d8dc9 100644 --- a/doc/merge.txt +++ b/doc/merge.txt @@ -20,6 +20,13 @@ The merge process - merge-dist distributes 'sth' and missing entries to frontend nodes. + [logorder] + v + fetch backup sth dist + v ^ v ^ v ^ + [fetched] [backup.<secondary>] [sth] + + Merge distribution (merge_dist) ----------------------------------------------------- @@ -83,8 +90,6 @@ Merge backup (merge_backup) TODO ==== -- Run the three pieces in parallell. - - Improve merge-fetch by parallellising it using one process per storage node writing to a "queue info" file (storage-node, hash) and a single "queue handling process" reading queue files and writing to the |