MogileFS Recent threads

Threads	Replies	First post	Last post
$have_dbh variable in monitor device checks By: Eric Wong Do we really need to keep this variable around all the way through? I think MogileFS::Store holds onto it for us so we won't get inadvertantly disconnected. Anyways, I started on refactoring monitor and rewriting it to use Danga::Socket for HTTP…	3	May 23 2012 17:53	May 23 2012 18:32
Limit disk usage By: Jon Skarpeteig Is there any way to keep a minimum amount of free space on a storage disk? (without breaking functionality) Ideally, this would be done by tracker and a config option. Can disk quotas be used? (Or will this conflict with space left on device…	2	May 23 2012 14:19	May 23 2012 14:35
Mogilef php extension By: ki0 Hi everyone, Im trying to install mogilefs-php-0.9.1 version on debian with php 5.3.13, and it seems it doesnt work any more, i got this message: pecl/mogilefs requires PHP (version >= 5.2, version <= 5.3.3), installed version is 5.3.13 …	1	May 21 2012 13:35	May 23 2012 14:15
cmogstored 0.2.0 released By: Eric Wong cmogstored is an _experimental_, alternative implementation of the "mogstored" storage component of MogileFS. Unlike MogileFS itself, cmogstored is not considered stable nor sane for production use. cmogstored is implemented in C and does not use…	5	Mar 17 2012 19:20	May 19 2012 19:48
::Server 2.62 going up By: dormando Hey, To fix a compilation issue, and an issue in upgrading Postgres to support checksums, I've made another release of ::Server. Much thanks to Robin for pointing out and patching the issues. Removing files from the dist is always tricky, I screw…	1	May 19 2012 15:58	May 19 2012 18:53
Update Pg locking code to use Postgres advisory locks. By: Robbat2 From: "Robin H. Johnson" <rob### @gentoo.org> Signed-off-by: Robin H. Johnson <robb### @gentoo.org> --- CHANGES \| 2 ++ lib/MogileFS/Store/Postgres.pm \| 38 +++++++++++++++++++++	10	May 18 2012 11:57	May 19 2012 16:01
MogileFS::Server 2.61 going up By: dormando Yo, I may have too aggressively cut the CHANGES file by accident :( This is a sweet bugfix and testing release, mostly driven by Eric's hard work! The last two cuts were pretty far apart, lets see if I can get it back down to one release every…	10	May 19 2012 01:37	May 19 2012 15:56
Postgres: Fix v15 schema upgrade. By: Robin H. Johnson From: "Robin H. Johnson" <robb### @gentoo.org> Schema upgrade needs to use Pg-specific column types for the v15 upgrade adding class.hashtype. Only CREATE TABLE is auto-converted where possible, not ALTER TABLE. Signed-off-by: Robin H.…	0	May 19 2012 15:00	May 19 2012 15:00
PATCH 0/2] dead code elimination branch By: Eric Wong While working on other stuff, I've also encountered some unused code. I'll continually update/rebase this branch against master: git://bogomips.org/MogileFS-Server deadcode If we need anything I've killed for future purposes or if there are…	22	May 4 2012 13:38	May 19 2012 00:40
mogdbsetup - plugins By: Piekarczyk Micha? Hi, i have found that once MogileFS is set up, a plugin is not able to create any new tables from within mongodbsetup. I attach a simple patch to fix this issue. Please let me know, what you think about it. Cheers	1	Apr 27 2012 04:12	May 18 2012 21:22
Better handling of mostly full devices. By: Tomas Doran The attached patch allows clients to (optionally) send the size of the file they are about to write to the tracker. This allows the tracker to filter out any devices which are known to not have enough space free for this write. Currently the…	7	Feb 28 2012 06:21	May 18 2012 21:14
Make valid_state Host check stricter By: Pyry Hakulinen Regex is wrong and all states starting with alive, or containing dead, or ending with down will match it. Fixes: http://code.google.com/p/mogilefs/issues/detail?id=57 --- lib/MogileFS/Host.pm \| 2 +- 1 files changed, 1 insertions(+), 1…	2	Mar 9 2012 06:38	May 18 2012 21:02
cmogstored 0.4.0 - stable for GNU/Linux users By: Eric Wong Summary of changes since cmogstored 0.3.0: * The kqueue code path now requires no more syscalls than the epoll code path on BSDs. avoids malloc() usage after fork() (for iostat(1)). This works around some malloc() implementations which…	0	May 18 2012 19:53	May 18 2012 19:53
util: every() responds to ":wake_a" commands By: Eric Wong Instead of blindly sleeping, workers select() on psock_fd in order to respond to ":wake_a" commands. This allows fsck to start immediately when triggered. This change does not increase wakeups during idle periods[1]. This fixes a regression…	7	May 12 2012 21:12	May 18 2012 19:52
Re: rfc json By: Gernot Vormayr > > - Test to be sure 'JSON' actually uses an ::XS backend if available, this > would let us reduce CPU usage while switching to JSON > I would print a warning if JSON::PP backend is used > - Run both protocol side by side…	4	Apr 30 2012 22:07	May 18 2012 14:06
Make it easier to get stats on rebalance during the test. By: Robbat2 From: "Robin H. Johnson" <robb### @gentoo.org> Signed-off-by: Robin H. Johnson <robb### @gentoo.org> --- t/30-rebalance.t \| 59 +++++++++++++++++++++++++++++++++++++++++	0	May 18 2012 11:57	May 18 2012 11:57
PATCH 0/2] sqlite3 locking implementation By: Eric Wong I've lightly tested this, seems to pass all tests. I mainly need this to have locks around the delete worker. I think the following is ready to pull, but an extra set of eyes would be appreciated (it's late and I'm sleepy): git pull…	17	Mar 1 2012 05:43	May 18 2012 11:05
remove redundant code to ignore SIGPIPE By: Eric Wong MogileFS::Server and Perlbal both ignore SIGPIPE for us. So there's no need to ever ignore it for socket writes in HTTPFile, either. --- This is against master. $ git pull git://bogomips.org/MogileFS-Server.git redundant-sigpipe MANIFEST …	1	May 11 2012 22:23	May 17 2012 22:50
fsck: prevent running over 100% completion By: Eric Wong FIDs may be created while fsck is running, causing "mogadm fsck status" to report completion above 100% (and thus confusing users). Stopping fsck when it reaches fsck_fid_at_end (set to MAX(fid) at fsck startup) prevents this. This change should…	1	May 17 2012 20:38	May 17 2012 21:26
tests: use done_testing() instead of test counts By: Eric Wong Keeping track of explicit test counts causes needless merge conflicts. done_testing() is sufficient to note test completeness and detect failures. --- I've also been using done_testing() in my fsck test. Some of my unpulled/RFC branches don't…	6	May 2 2012 20:54	May 17 2012 21:01