Created
January 4, 2017 23:11
-
-
Save pauldardeau/0081c7ccf2e58ad4d52fd609a9051cd0 to your computer and use it in GitHub Desktop.
all your loopback device space r belong to swift mp3
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
this is related to my dinky cloud jukebox (https://github.com/pauldardeau/cloud-jukebox). that's a fancy | |
way of saying a bunch of mp3 files stored in swift. | |
i run it on a vps (droplet!) with 2gb ram and 40gb disk | |
i continue to work on my cloud jukebox thingie | |
primarily by adding more of my music collection to my little saio swift cluster on a vps | |
as part of my goal of moving completely away from itunes | |
i’ve been doing a lot more of that lately, and i completely filled up (100%) my loopback storage device | |
it’s configured with standard 3 replicas (default) | |
i thought: “well, i don’t need 3 copies" | |
i’ll just adjust to 2 and that’ll free up some space | |
i set it to 2 and issue rebalance command | |
which tells me that it refuses to rebalance because it would only affect some small percent | |
i spent some time digging up the ‘force’ option | |
so i proudly copied my updated ring file (now with 2 replicas) to my saio | |
then i realized that i would need the replicator to move things around | |
(and it wasn’t already running) | |
i started the replicator, and see it cause bursts of cpu and disk usage | |
still 100% full | |
wait some more | |
more bursts | |
still 100% full | |
wait some more | |
more bursts | |
still 100% full | |
you get the idea... | |
then i went hunting down the log files | |
tail/follow one of the log files and see that rsync is failing | |
oh, yeah. rsync! | |
it’s not running — not even configured! | |
configure rsync, start it | |
wait a bit | |
more bursts | |
still 100% full | |
me thinks: maybe i should have taken a longer vacation | |
look around more log files | |
oh, maybe the replication needs a little bit of free space to do its thing | |
delete all *.data files on one of my virtual devices (loopback device on saio) | |
down from 100% full!!!! | |
wait a bit | |
more bursts | |
df -k showing differing values. might be working! | |
wait some more | |
more bursts | |
df -k settles down | |
decide it’s probably done moving things around | |
stop object-replicator (my little vps is already tight on resources) | |
65% space used. yay! | |
all of this happened on my super-tiny vps with 2 gb ram while it stayed running | |
and me still listening to music from it!!! | |
cheers to swift!!! | |
this was my “oh, it’ll only take a minute exercise” for today | |
my background task to enhance my music listening pleasure while doing my 'real work' | |
kinda embarrassed it took me so long to fix | |
but i didn't sleep enough last night | |
and i'm still getting my brain put in gear after the holidays | |
or at least that's my excuse (and sticking with it)! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment