#ceph IRC Log

Index

IRC Log for 2010-12-18

Timestamps are in GMT/BST.

[0:02] * midnightmagic (~nouser@S0106000102ec26fe.gv.shawcable.net) has joined #ceph
[0:02] <midnightmagic> Out of curiosity, one question the about page doesn't seem to answer is whether you can have programmable reduncancy? Say, via erasure-coded blocks or just simple X-copies redundancy amongst storage nodes? (Really like the idea of auto-balancing by the way.. soo.. awesome..)
[0:07] <sage> we only support n-way replication
[0:13] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) Quit (Quit: Yoric)
[0:23] <midnightmagic> any plans on k-of-n erasure-coded replication to reduce per-storage node storage requirements?
[0:23] <sage> not in the forseeable future :)
[0:24] <midnightmagic> assume I am a competent coder. can you see technical hurdles that would make k-of-n erasure-coding an infeasible goal? infrastructure assumptions/protocol requirements/etc?
[0:25] <midnightmagic> sorry.. let me be more accurate "an infeasible short-term goal" :-)
[0:25] <sage> the object storage is already very complex with just replication.. doing k-of-n is a pretty big thing. it would likely change the client protocol, the replication code, the recovery.. lots and lots on the backend especially
[0:26] <ijuz__> does ceph try to have copies close at the user and/or is there a notion about the connection speed to the OSD nodes?
[0:28] <gregaf> ijuz__: nope, Ceph basically assumes a single data center with essentially the same latency to all nodes
[0:29] <sage> no. not unless the administrator structures the crush mapping appropriately (e.g. with faster "front-line" nodes and slower backside replica nodes)
[0:29] <ijuz__> ok thanks, could have been, would be interesting for high data rates or big networks
[0:29] <gregaf> you can massage it a bit via CRUSH maps and stuff to have a sort of "standby" data center that doesn't serve writes if you've got sufficiently fast interconnects, but there's no way to make it adaptive based on client location/ping times
[0:35] <ijuz__> having offsite backup nodes would be really great, Coda has (or had) such a feature
[0:36] <cmccabe1> AFS also had a way to create read-only replicas
[0:37] <gregaf> offsite backup nodes require a full bandwidth interconnect to make work in Ceph, though :(
[0:37] <midnightmagic> sage: thank you, I really appreciate your time. :) got a bitcoin address I can use to donate to the project?
[0:39] <gregaf> people who want AFS-style replicas and stuff I tend to point towards extremeFS, as it seems to be trying to do that
[0:40] <sage> there's no way to donate currently. we tend to be short on people/time/testers and not funds, though (yay us)
[0:42] <iggy> and if you know anything about either of those filesystems, you'll know why it's not a high priority for all of the modern distributed network fses to implement
[0:42] <midnightmagic> ok
[0:42] <iggy> (i.e. they are both crap)
[0:43] <iggy> besides, in linux, there's enough kernel support for hooking into fses to do it all in userspace
[0:43] <iggy> well, at least the replication part, maybe not the locality part
[1:34] * bchrisman (~Adium@70-35-37-146.static.wiline.com) Quit (Quit: Leaving.)
[3:28] * joshd (~joshd@ip-66-33-206-8.dreamhost.com) Quit (Quit: Leaving.)
[3:35] * sjust (~sam@ip-66-33-206-8.dreamhost.com) Quit (Ping timeout: 480 seconds)
[3:55] * eternaleye_ is now known as eternaleye
[3:55] * cmccabe1 (~cmccabe@adsl-75-37-24-225.dsl.pltn13.sbcglobal.net) Quit (Quit: Leaving.)
[5:11] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) has joined #ceph
[6:44] * ijuz__ (~ijuz@p4FFF724B.dip.t-dialin.net) Quit (Ping timeout: 480 seconds)
[6:54] * ijuz__ (~ijuz@p4FFF4EF4.dip.t-dialin.net) has joined #ceph
[9:04] * allsystemsarego (~allsystem@188.25.128.213) has joined #ceph
[10:03] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[10:14] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) has joined #ceph
[10:35] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[11:08] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) has joined #ceph
[12:00] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[14:03] * votz (~votz@dhcp0020.grt.resnet.group.upenn.edu) has joined #ceph
[14:15] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[15:19] * ghaskins_mobile (~ghaskins_@66-189-113-47.dhcp.oxfr.ma.charter.com) has joined #ceph
[15:19] * ghaskins_mobile (~ghaskins_@66-189-113-47.dhcp.oxfr.ma.charter.com) Quit ()
[16:06] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) Quit (Quit: Yoric)
[16:30] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) has joined #ceph
[16:38] * MarkN (~nathan@mail.zomojo.com) has joined #ceph
[16:59] * shdb (~shdb@217-162-231-62.dclient.hispeed.ch) Quit (Read error: Connection reset by peer)
[17:07] * shdb (~shdb@217-162-231-62.dclient.hispeed.ch) has joined #ceph
[17:46] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[18:19] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[18:29] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[18:39] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[18:40] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[18:47] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[18:48] * bchrisman (~Adium@c-24-130-226-22.hsd1.ca.comcast.net) has joined #ceph
[19:03] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[19:04] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[19:13] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[19:21] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[19:41] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[19:41] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[19:49] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[19:50] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) has joined #ceph
[20:02] * tjikkun (~tjikkun@2001:7b8:356:0:204:bff:fe80:8080) Quit (Ping timeout: 480 seconds)
[22:16] * allsystemsarego (~allsystem@188.25.128.213) Quit (Quit: Leaving)

These logs were automatically created by CephLogBot on irc.oftc.net using the Java IRC LogBot.