Context Navigation

01:38 Changeset [699] by hachi: r23755@lj: lj | 2007-01-10 22:08:52 -0800 Add utilization to the ...
01:38 Changeset [698] by hachi: r23730@lj: lj | 2007-01-10 17:45:20 -0800 This should fix a warning ...

01/10/07:

00:29 Changeset [697] by jay: clarified message whenever DBD::MySQL isn't installed and user doesn't set ...

01/09/07:

04:14 Changeset [696] by hachi: r23652@lj: lj | 2007-01-08 19:14:20 -0800 Unused code... as far as I ...
04:13 Changeset [695] by hachi: r23649@lj: lj | 2007-01-08 19:10:16 -0800 Switch all logic for ...
04:13 Changeset [694] by hachi: r23648@lj: lj | 2007-01-08 19:09:18 -0800 Syntax error.
03:17 Changeset [693] by hachi: r23643@lj: lj | 2007-01-08 18:16:35 -0800 I changed the meaning of the ...
03:16 Changeset [692] by botemployee: finally adding this
02:37 Changeset [691] by bradfitz: dear svk users, plz not to be backing out my patches, thx. :)
02:14 Changeset [690] by hachi: r23620@lj: lj | 2007-01-08 17:13:38 -0800 Comments for clarity
02:06 Changeset [689] by hachi: r23615@lj: lj | 2007-01-08 17:06:01 -0800 Working device weighting ...
02:06 Changeset [688] by hachi: r23614@lj: lj | 2007-01-08 17:05:04 -0800 Working test to exersize the ...
02:06 Changeset [687] by hachi: r23613@lj: lj | 2007-01-08 16:21:05 -0800 This should be used in here ...
02:06 Changeset [686] by hachi: r23612@lj: lj | 2007-01-08 16:20:07 -0800 Make it run under the current ...
01:08 Changeset [685] by hachi: r23610@lj: lj | 2007-01-08 16:07:42 -0800 Non-working test intended for ...

01/08/07:

20:39 Changeset [684] by bradfitz

work better with sqlite? still can't fix db locking issue.

19:55 Changeset [683] by hachi

r23590@lj: lj | 2007-01-08 10:25:59 -0800 Propagate device utilization ...

19:55 Changeset [682] by hachi

r23536@lj: lj | 2007-01-05 17:56:02 -0800 Set the utilization on the ...

19:55 Changeset [681] by hachi

r23535@lj: lj | 2007-01-05 17:38:47 -0800 Switch to displaying just ...

19:55 Changeset [680] by hachi

r23534@lj: lj | 2007-01-05 16:42:00 -0800 Docs for the IOStatWatcher ...

19:55 Changeset [679] by hachi

r23533@lj: lj | 2007-01-05 16:27:22 -0800 Disks returns a count of how ...

19:55 Changeset [678] by hachi

r23532@lj: lj | 2007-01-05 15:52:18 -0800 Fix the iostat watcher so it ...

16:36 Changeset [677] by bradfitz

insert_ignore wrapper for databases without it

16:26 Changeset [676] by bradfitz

more db portability work. sqlite getting further.

16:05 Changeset [675] by bradfitz

move register_tempfile up into base class, making it more portable ...

15:19 Changeset [674] by bradfitz

push 3 more functions up into base class, using abstract unix_timestamp

15:06 Changeset [673] by bradfitz

move 3 functions up from MySQL/SQLite into common base class. they used ...

14:58 Changeset [672] by bradfitz

ask the right class for whether it wants RaiseError

14:40 Changeset [671] by bradfitz

abstract out UNIX_TIMESTAMP().

also don't use mysql-ish SET= syntax on ...

14:31 Changeset [670] by bradfitz

load the Store base class

00:18 Changeset [669] by bradfitz

this makes 3 of the tests pass now for sqlite

01/07/07:

23:10 Changeset [668] by bradfitz: more sqlit work
23:04 Changeset [667] by bradfitz: sqlite now passes store.t (which is admittely pretty small, but it's a ...
22:35 Changeset [666] by bradfitz: start of SQLite support. schema creation now passes, and tests ...
21:38 Changeset [665] by bradfitz: map from oracle/postgres DSNs to their class names.
21:35 Changeset [664] by bradfitz: finish moving all of mogdbsetup into the Store classes
20:31 Changeset [663] by bradfitz: start making mogdbsetup use the Store interface for making new ...
19:05 Changeset [662] by bradfitz: remove mentions of mysql from t/*
04:16 Changeset [661] by bradfitz: remove TempDB crap in mogtestlib.pl, instead relying on the Store to be ...

01/06/07:

05:32 Changeset [660] by bradfitz

start converting Worker/Delete to use Store

02:48 Changeset [659] by bradfitz

prep work to start cleaning up Worker/Delete

02:48 Changeset [658] by bradfitz

remove debug

01:44 Changeset [657] by bradfitz

remove final MySQL-ism from Worker/Replicate.

01:03 Changeset [656] by bradfitz

Internal protocol/APIs for waking up sleeping worker processes.

For ...

01/05/07:

23:18 Changeset [655] by bradfitz

add delete tests

22:39 Changeset [654] by bradfitz

* Improve lock contention between replicate processes on the same box by ...

21:44 Changeset [653] by bradfitz

minor replication cleanup, mostly comments, and update the ...

20:53 Changeset [652] by bradfitz

fix --daemonize

20:50 Changeset [651] by bradfitz

more js

20:37 Changeset [650] by bradfitz

memcache notes

20:22 Changeset [649] by bradfitz

to exclude brad's symlinks

20:13 Changeset [648] by bradfitz

replication/class FIXME cleanup.... fids should always have classes, even ...

12:39 Changeset [647] by hachi

r23524@lj: lj | 2007-01-05 03:38:33 -0800 Switch get_keys command to ...

10:49 Changeset [646] by hachi

10:44 Changeset [645] by hachi

r23461@lj: lj | 2007-01-04 16:54:32 -0800 Perform list_fids from the ...

10:44 Changeset [644] by hachi

r23460@lj: lj | 2007-01-04 15:51:37 -0800 Better doc generation... now ...

08:45 Changeset [643] by bradfitz

just about done cleaning up Worker/Replicate to be dbh-free

07:54 Changeset [642] by bradfitz

cleaning up DBI stuff in Worker/Replicate

07:27 Changeset [641] by bradfitz

flesh out Host: Host->create and a bunch of setters.

removes one of the ...

05:46 Changeset [640] by bradfitz

remove last dbh from Domain

05:37 Changeset [639] by bradfitz

remove last dbh from Device

00:42 Changeset [638] by bradfitz

forgot best part: removing dbh!

00:41 Changeset [637] by bradfitz

add MogileFS::Device->create method. more cleanup.

00:36 Changeset [636] by bradfitz

start of new tests

00:12 Changeset [635] by bradfitz

remove another dbh

00:10 Changeset [634] by bradfitz

remove more $dbh from query worker

00:00 Changeset [633] by bradfitz

shitload of work that spiraled out of control once I started ...

01/04/07:

21:25 Changeset [632] by bradfitz

more domain testing work.

21:25 Changeset [631] by bradfitz

TODO

21:04 Changeset [630] by bradfitz

rename classes.t to domains-classes.t, add a bunch of domain tests, add an ...

20:34 Changeset [629] by hachi

r23452@lj: lj | 2007-01-04 11:30:09 -0800 Switch update_devcount to use ...

20:34 Changeset [628] by hachi

r23451@lj: lj | 2007-01-04 11:14:27 -0800 Enqueue for replication using ...

20:27 Changeset [627] by bradfitz

misc forgotten things from previous commit

20:27 Changeset [626] by bradfitz

make MogileFS::Server testable, without starting server. and start of ...

20:25 Changeset [625] by bradfitz

todo

20:10 Changeset [624] by bradfitz

add MogileFS::Server, for server singleton.

now mogilefsd is just:

...

20:08 Changeset [623] by bradfitz

put these uses here.

19:55 Changeset [622] by hachi

r23444@lj: lj | 2007-01-04 10:37:45 -0800 Move enumeration of fidids ...

19:55 Changeset [621] by hachi

r23443@lj: lj | 2007-01-04 10:22:33 -0800 Move the ...

19:55 Changeset [620] by hachi

r23442@lj: lj | 2007-01-04 10:18:28 -0800 Comments

19:55 Changeset [619] by hachi

r23441@lj: lj | 2007-01-04 10:15:42 -0800 Don't implicitly set the ...

19:43 Changeset [618] by bradfitz

use the constant for this that already existed

08:37 Changeset [617] by bradfitz

cleanpu

08:21 Changeset [616] by bradfitz

catch exceptions in queryworker and return them to client as errors.

07:17 Changeset [615] by bradfitz

update cmd_delete_host

07:00 Changeset [614] by bradfitz

hacking status tool

06:39 Changeset [613] by bradfitz

more cleanup.

06:28 Changeset [612] by bradfitz

make MogileFS::Domain an object, and make Worker/Query start to use it

05:49 Changeset [611] by bradfitz

start reworking MogileFS::Domain, (later to be) turning it into a ...

05:01 Changeset [610] by hachi

Add update_devcount methods to the store, not ready to be used yet because ...

03:53 Changeset [609] by bradfitz

make query worker's create_class (which also updates classes) use the ...

03:24 Changeset [608] by bradfitz

remove DBI stuff from Host

03:17 Changeset [607] by bradfitz

use the Store::MySQL get_lock/release_lock wrappers that check if we're ...

03:00 Changeset [606] by hachi

Get all devices from within the store.

02:57 Changeset [605] by hachi

Add safe get_lock and release_lock methods for the MySQL Store. Brad, you ...

01:37 Changeset [604] by hachi

Remove another dbh usage.

01/03/07:

23:38 Changeset [603] by hachi

Better comments, and don't use the deprecated API for getting the dbh.

23:31 Changeset [602] by hachi

Forgot to delete debug line before comitting.

23:28 Changeset [601] by hachi

Move dbh usage into Store.

I decided to use map in Store.pm sub ...

20:25 Changeset [600] by bradfitz

work in dbi-not-RaiseError'ing case. will be cleaner later.

01/02/07:

22:47 Changeset [599] by bradfitz: fix bug in argument name, preventing device times from ever being ...
22:47 Changeset [598] by bradfitz: invalidate cache correctly with new APIs, not old missing ones
22:47 Changeset [597] by bradfitz: use shared code
20:48 Changeset [596] by bradfitz: was nexting out of a subrouting.
02:30 Changeset [595] by bradfitz: I'd added a duplicate method.
02:24 Changeset [594] by bradfitz: modern cmd_rename
02:08 Changeset [593] by bradfitz: finish making cmd_create_close clean
01:11 Changeset [592] by bradfitz: start making cmd_create_close be more OO and less dbh-using
00:39 Changeset [591] by bradfitz: more modernization, now of cmd_set_state, and update cmd_set_weight more, ...

01/01/07:

23:40 Changeset [590] by bradfitz: modernize cmd_delete.
23:28 Changeset [589] by bradfitz: make cmd_set_weight be OO and use the store, not dbh
23:12 Changeset [588] by bradfitz: caps

12/29/06:

08:48 Changeset [587] by bradfitz

doc fixes

06:52 Changeset [586] by bradfitz

more docs

06:48 Changeset [585] by bradfitz

had forgot to commit this... more get_dbh removal

06:46 Changeset [584] by bradfitz

had forgot to commit this... more get_dbh removal

06:43 Changeset [583] by bradfitz

start of docs

06:21 Changeset [582] by bradfitz

useful dev tool

06:03 Changeset [581] by bradfitz

tagging 04

06:02 Changeset [580] by bradfitz

dev version 04 snapshot

05:29 Changeset [579] by bradfitz

get_dbh() removal

05:21 Changeset [578] by bradfitz

remove get_dbh caller

05:19 Changeset [577] by bradfitz

move random add_file_on() function, previously stranded in the replication ...

05:06 Changeset [576] by bradfitz

don't use dbh, use the store.

04:56 Changeset [575] by bradfitz

remove unused stuff, use debug

04:55 Changeset [574] by bradfitz

remove unused stuff. this whole file is unused. what is diff between ...

03:17 Changeset [573] by bradfitz

moving more to store, removing dbh users

03:02 Changeset [572] by bradfitz

move all the mysql tempfile bullshit into MogileFS::Store::MySQL, making ...

02:33 Changeset [571] by bradfitz

make the old Mgd::get_dbh() and Mgd::validate_dbh() just wrappers around ...

02:32 Changeset [570] by bradfitz

don't re-ping dbh always, only when marked as needs pinging (and add new ...

02:31 Changeset [569] by bradfitz

make pidfile default to "", not undef, so it's defined

02:14 Changeset [568] by bradfitz

pidfile support

01:26 Changeset [567] by bradfitz

2 get_dbh() down, 54 to go....

01:25 Changeset [566] by bradfitz

increase timeout to 10 seconds for monitor, not the default of 3 from the ...

01:16 Changeset [565] by bradfitz

start of database-independence work.

introduces MogileFS::Store::*

01:15 Changeset [564] by bradfitz

whitespace only

00:23 Changeset [563] by bradfitz

unindent

00:22 Changeset [562] by bradfitz

fix the broken-earlier-today altip/altzone code, removing yet more code ...

12/28/06:

23:52 Changeset [561] by bradfitz

move server start time (and new accessor) into ProcManager, and out of ...

23:43 Changeset [560] by bradfitz

remove Mgd::close_listeners, and instead register with procmanager a ...

23:35 Changeset [559] by bradfitz

move server_setting and schema checking stuff into MogileFS::Config

23:26 Changeset [558] by bradfitz

-- Finish killing Mgd::get_device_summary(), from everywhere.

-- flesh ...

21:43 Changeset [557] by bradfitz

this removes all callers of Mgd::get_device_summary() that are NOT in ...

21:27 Changeset [556] by bradfitz

start to get rid of callers of Mgd::get_device_summary.

21:00 Changeset [555] by bradfitz

Introduce MogileFS::DevFID, which represents a (device,fid) tuple. then ...

19:54 Changeset [554] by bradfitz

move find_deviceid into MogileFS::Device for now, even though that's not ...

19:48 Changeset [553] by bradfitz

move all host caching/loading out of mogilefsd and into MogileFS::Host

18:43 Changeset [552] by bradfitz

from Mgd::invalidate_class_cache() to MogileFS::Class->invalidate_cache

18:29 Changeset [551] by bradfitz

move Mgd::class_name()

18:11 Changeset [550] by bradfitz

move vivify_directories into MogileFS::Device

17:59 Changeset [549] by bradfitz

move all the observed_foo APIs into MogileFS::Host and MogileFS::Device, ...

05:39 Changeset [548] by bradfitz

move all domain-related functions into MogileFS::Domain and out of ...

05:34 Changeset [547] by bradfitz

move Mgd::domain_id() into new file, MogileFS::Domain (which isn't yet an ...

05:10 Changeset [546] by bradfitz

get rid of Mgd::key_filerow. move it into a new ctor for MogileFS::FID ...

04:57 Changeset [545] by bradfitz

move is_url out of mogilefsd and into MogileFS::Util, renamed url_parts

04:52 Changeset [544] by bradfitz

move Mgd::hostid_classes() and its one caller to ...

04:49 Changeset [543] by bradfitz

faster file check logic

04:38 Changeset [542] by bradfitz

move class_id into MogileFS::Class

04:27 Changeset [541] by bradfitz

move wait_for_readability and wait_for_writeability into MogileFS::Util

04:18 Changeset [540] by bradfitz

move Mgd::get_mindevcounts() into MogileFS::Class

04:14 Changeset [539] by bradfitz

more kill old APIs. this time Mgd::update_fid_devcount();

03:53 Changeset [538] by bradfitz

more legacy API removal. this time killing Mgd::host_id()

03:42 Changeset [537] by bradfitz

converting more code to OO API, killing old functions in Mgd:: in ...

03:34 Changeset [536] by bradfitz

remove sleep at end, and fix test count

03:28 Changeset [535] by bradfitz

variable renaming only, in prep for future patch

03:26 Changeset [534] by bradfitz

remove some non-OO Mgd:: code

03:22 Changeset [533] by bradfitz

this wasn't used

03:21 Changeset [532] by bradfitz

more work on and usage of Mogile::Host.

plan is to slowly drain the ...

02:41 Changeset [531] by bradfitz

use keep alives from monitor

02:13 Changeset [530] by bradfitz

make the monitor code even more object-oriented, this time focusing on the ...

01:32 Changeset [529] by bradfitz

update the monitor worker to be cleaner, smaller, and use the ...

01:15 Changeset [528] by bradfitz

class method to return all devices, as well as new accessor to see if ...

01:14 Changeset [527] by bradfitz

Mgd::error() from MogileFS::Util, used to be used for errors, fatal things ...

12/27/06:

23:45 Changeset [526] by bradfitz: much expanded tests
23:34 Changeset [525] by bradfitz: don't send status updates, only in debug mode.
22:26 Changeset [524] by bradfitz: require Danga::Socket 1.56 which resets epoll stuff for monitor worker.
21:47 Changeset [523] by bradfitz: the real bug fix: can't "next" from a sub. well, you can... perl ...
21:46 Changeset [522] by bradfitz: die hard if a child process is getting a message from a child process. ...
21:46 Changeset [521] by bradfitz: add a more-several error() wrapper: fatal.
21:45 Changeset [520] by bradfitz: * kill child processes (diskusage and iostat) when being killed.
21:44 Changeset [519] by bradfitz: unused

Note: See TracTimeline for information about the timeline view.

Download in other formats:

RSS Feed