#ceph IRC Log


IRC Log for 2010-07-30

Timestamps are in GMT/BST.

[0:10] * pruby (~tim@leibniz.catalyst.net.nz) Quit (Remote host closed the connection)
[0:11] * pruby (~tim@leibniz.catalyst.net.nz) has joined #ceph
[0:16] * allsystemsarego (~allsystem@ Quit (Quit: Leaving)
[0:50] * anthony (~anthony@dhcp-v021-174.mobile.uci.edu) Quit (Read error: Connection reset by peer)
[0:57] * eternaleye (~quassel@184-76-53-210.war.clearwire-wmx.net) Quit (Ping timeout: 480 seconds)
[1:03] * eternaleye (~quassel@184-76-53-210.war.clearwire-wmx.net) has joined #ceph
[1:24] * anthony (~anthony@dhcp-v021-174.mobile.uci.edu) has joined #ceph
[1:25] <anthony> hi, i have a question about ceph crushmaps
[1:25] <anthony> can anyone help? =/
[1:25] <sagewk> whats the question?
[1:25] <anthony> http://pastebin.com/pPwqVwn1
[1:26] <anthony> the map that i made isn't working, i can't even mount ceph after i set the new crush map so i was wondering if something is wrong with the configuration
[1:26] <sagewk> change "choose" in teh rules to "choose_leaf"
[1:27] <sagewk> also, i _think_ you can leave the 'weight 4.0" and 8.0 off in the root root {} section and it'll auto calculate it when the map compiles. or leave it, doesn't matter.
[1:27] <anthony> ahh.
[1:27] <anthony> okay
[1:27] <anthony> thanks sage
[2:09] * anthony (~anthony@dhcp-v021-174.mobile.uci.edu) Quit (Quit: Leaving)
[2:52] * akhurana (~ak2@c-98-232-30-233.hsd1.wa.comcast.net) has joined #ceph
[2:53] * akhurana (~ak2@c-98-232-30-233.hsd1.wa.comcast.net) Quit ()
[5:21] * ajnelson (~ajnelson@wsip-24-234-51-36.lv.lv.cox.net) has joined #ceph
[5:22] * ajnelson (~ajnelson@wsip-24-234-51-36.lv.lv.cox.net) Quit ()
[6:08] * Osso (osso@AMontsouris-755-1-7-241.w86-212.abo.wanadoo.fr) Quit (Quit: Osso)
[6:51] * f4m8_ is now known as f4m8
[8:31] * eternaleye (~quassel@184-76-53-210.war.clearwire-wmx.net) Quit (Ping timeout: 480 seconds)
[10:22] * allsystemsarego (~allsystem@ has joined #ceph
[15:00] * f4m8 is now known as f4m8_
[15:25] * Osso (osso@AMontsouris-755-1-7-241.w86-212.abo.wanadoo.fr) has joined #ceph
[18:57] <wido> hi
[18:57] <wido> has there been a monmap change since yesterday? I saw a commit about it, but i'm not sure
[18:57] <wido> my mon's refuse to start with the latest unstable, they complain that they are not present in the monmap
[19:01] <sagewk> yeah, the mon-remove branch was merged. it should start, though.. i'll take a look
[19:02] <wido> sagewk: http://pastebin.org/433550
[19:02] <wido> i tried to create a new monmap
[19:08] <sagewk> hmm, i think you must have briefly run an unstable from yesterday that had the monitor decoding goofed up.
[19:09] <wido> could i create a new monmap?
[19:09] <sagewk> i switched back to latest.bak and renamed your montiors mon.a and mon.b, and it's up. you'll probably want to resync your ceph.conf across the cluster
[19:10] <wido> ah, ok :)
[19:10] <wido> i'll do so
[19:10] <sagewk> yeah.. but it can't be copied to monmap/latest verbatim. there is a u64 version and u32 length prefix
[19:10] <sagewk> it's only because you ran the interim broken version.. moving from 0.21 to current unstable shouldn't break.
[19:11] <sagewk> (or required the conf change)
[19:11] <wido> so my mon's are named mon.a and mon.b now
[19:11] <sagewk> yeah
[19:11] * eternaleye (~quassel@184-76-53-210.war.clearwire-wmx.net) has joined #ceph
[19:11] <sagewk> the new scheme gives them proper names, so they can be added/removed without the numbering getting screwed up
[19:11] <sagewk> similar to the mds's.
[19:12] <wido> ah, i get it :)
[19:12] <sagewk> the osds still get numeric names, since their rank is directly tied to what data they store (unlike the mds (no local data) and monitor (all store identical data))
[19:13] <wido> yes, i get it
[19:13] <wido> i'll manage
[19:13] <wido> have to go afk
[20:23] * allsystemsarego (~allsystem@ Quit (Ping timeout: 480 seconds)
[21:30] * allsystemsarego (~allsystem@ has joined #ceph
[23:03] * allsystemsarego (~allsystem@ Quit (Quit: Leaving)
[23:13] * ghaskins_mobile (~ghaskins_@66-189-114-103.dhcp.oxfr.ma.charter.com) Quit (Quit: This computer has gone to sleep)
[23:51] <wido> yehudasa: you there?
[23:51] <yehudasa> yeah
[23:52] <wido> i've been testing today with the rgw, but changing the acl of a bucket fails.
[23:52] <wido> the new XML is generated and then run through the decode method, but the xattr of the rados pool isn't changed
[23:52] <wido> while in rgw_rados.cc the setxattr method is called
[23:53] <yehudasa> hmm
[23:53] <yehudasa> yeah.. I'm not sure it's supposed to work
[23:54] <wido> i don't know what Amazon does, let me check that out
[23:55] <gregaf> do you mean that you can't give permission to other rgw users to access it?
[23:55] <wido> no, i created a bucket, but then i try to give All Users READ permission, that fails
[23:56] <wido> or other way around, created a bucket with the cannedAcl public-read, but then revoke the ACL on that bucket
[23:56] <wido> you can't change the ACL of a bucket after it's created
[23:58] * gregaf (~Adium@ip-66-33-206-8.dreamhost.com) Quit (Quit: Leaving.)
[23:58] <yehudasa> yeah, looking at it now
[23:58] * gregaf (~Adium@ip-66-33-206-8.dreamhost.com) has joined #ceph

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