#ceph IRC Log

Index

IRC Log for 2012-07-02

Timestamps are in GMT/BST.

[0:04] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) has joined #ceph
[0:41] * MarkDude (~MT@c-71-198-138-155.hsd1.ca.comcast.net) has joined #ceph
[1:13] * BManojlovic (~steki@212.200.241.207) Quit (Quit: Ja odoh a vi sta 'ocete...)
[1:54] * yoshi (~yoshi@p22043-ipngn1701marunouchi.tokyo.ocn.ne.jp) has joined #ceph
[2:17] * lofejndif (~lsqavnbok@9YYAAHRYV.tor-irc.dnsbl.oftc.net) Quit (Quit: gone)
[3:01] * renzhi (~renzhi@208.53.156.36) has joined #ceph
[3:36] * asadpanda (~asadpanda@2001:470:c09d:0:20c:29ff:fe4e:a66) Quit (Remote host closed the connection)
[4:39] * brambles (brambles@79.133.200.49) Quit (Remote host closed the connection)
[4:39] * brambles (brambles@79.133.200.49) has joined #ceph
[4:44] * brambles_ (brambles@79.133.200.49) has joined #ceph
[5:23] * asadpanda (~asadpanda@2001:470:c09d:0:20c:29ff:fe4e:a66) has joined #ceph
[6:02] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) Quit (Remote host closed the connection)
[6:33] * s[X] (~sX]@49.176.100.108) has joined #ceph
[6:34] * s[X] (~sX]@49.176.100.108) Quit (Remote host closed the connection)
[6:38] * f4m8 (f4m8@kudu.in-berlin.de) has joined #ceph
[6:48] * s[X] (~sX]@49.176.65.163) has joined #ceph
[7:07] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[7:09] * loicd (~loic@magenta.dachary.org) has joined #ceph
[7:12] * s[X] (~sX]@49.176.65.163) Quit (Remote host closed the connection)
[7:14] * widodh_ (~widodh@minotaur.apache.org) has joined #ceph
[7:20] * widodh (~widodh@minotaur.apache.org) Quit (Ping timeout: 480 seconds)
[7:45] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[7:47] * loicd (~loic@magenta.dachary.org) has joined #ceph
[8:11] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[8:12] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) has joined #ceph
[8:18] * hijacker (~hijacker@213.91.163.5) Quit (Quit: Leaving)
[8:21] * ninkotech (~duplo@89.177.137.231) Quit (Read error: Connection reset by peer)
[8:22] * ninkotech (~duplo@89.177.137.231) has joined #ceph
[8:49] * MarkDude (~MT@c-71-198-138-155.hsd1.ca.comcast.net) Quit (Quit: Leaving)
[8:52] * hijacker (~hijacker@213.91.163.5) has joined #ceph
[8:55] * Ryan_Lane (~Adium@c-98-210-205-93.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[8:55] * gregorg_taf (~Greg@78.155.152.6) Quit (Quit: Quitte)
[8:57] * verwilst (~verwilst@d5152FEFB.static.telenet.be) has joined #ceph
[9:04] * The_Bishop (~bishop@2a01:198:2ee:0:4122:7a01:9e5d:a950) Quit (Quit: Wer zum Teufel ist dieser Peer? Wenn ich den erwische dann werde ich ihm mal die Verbindung resetten!)
[9:09] * fghaas (~florian@91.119.51.139) has joined #ceph
[9:20] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[9:23] * loicd (~loic@83.167.43.235) has joined #ceph
[9:34] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) Quit (Remote host closed the connection)
[10:23] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[10:31] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[10:32] * ThoughtCoder (ThoughtCod@60-240-78-43.static.tpgi.com.au) has joined #ceph
[10:33] <ThoughtCoder> Hey guys, looking to add my ceph cluster to my fstab and was referenced line: "10.112.9.11,10.112.9.12,10.112.9.13:6789:/ /mnt/ceph ceph name=admin,secret=AQBBg0NPeFt9MRAAg8AWG108I8WDKHOWHdjkiw== POOLNAME,noauto,noatime 0 0" as an example - is the "POOLNAME" switch legitimate? Can I specify "rbd" or "data" here to specifically have that pool mount at boot?
[10:37] * liiwi (liiwi@idle.fi) Quit (Remote host closed the connection)
[10:40] * gregorg (~Greg@78.155.152.6) has joined #ceph
[10:51] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[11:08] * liiwi (liiwi@idle.fi) has joined #ceph
[11:11] * loicd1 (~loic@83.167.43.235) has joined #ceph
[11:11] * loicd (~loic@83.167.43.235) Quit (Read error: Connection reset by peer)
[11:14] * brambles_ (brambles@79.133.200.49) Quit (Quit: leaving)
[11:18] * ThoughtCoder (ThoughtCod@60-240-78-43.static.tpgi.com.au) Quit (Ping timeout: 480 seconds)
[11:30] * yoshi (~yoshi@p22043-ipngn1701marunouchi.tokyo.ocn.ne.jp) Quit (Remote host closed the connection)
[11:33] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Quit: Leaving.)
[11:37] * The_Bishop (~bishop@2a01:198:2ee:0:b44f:96e3:4d1f:3247) has joined #ceph
[11:40] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[11:51] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[12:38] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[12:44] * loicd1 is now known as loicd
[13:01] * gregorg_taf (~Greg@78.155.152.6) has joined #ceph
[13:01] * gregorg (~Greg@78.155.152.6) Quit (Read error: Connection reset by peer)
[13:05] * nhorman (~nhorman@hmsreliant.think-freely.org) has joined #ceph
[13:23] * widodh (~widodh@minotaur.apache.org) has joined #ceph
[13:23] * widodh_ (~widodh@minotaur.apache.org) Quit (Ping timeout: 480 seconds)
[14:35] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[15:32] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[15:57] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[16:11] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[16:32] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[16:55] <elder> sage?
[16:57] * elder (~elder@c-71-195-31-37.hsd1.mn.comcast.net) Quit (Quit: Leaving)
[16:57] * elder (~elder@c-71-195-31-37.hsd1.mn.comcast.net) has joined #ceph
[17:03] <sage> elder: here!
[17:03] <elder> OK.
[17:03] <elder> I was talking with Josh on Friday.
[17:03] <elder> I got my new-style requests to get to the server and back, but was getting various errors.
[17:04] <elder> I was sending a "get_immutable_metadata" request over, but it was not applying to the right object.
[17:05] <sage> not going to rbd_header.foo?
[17:05] <elder> Josh said that I needed to do the individual object requests rather than the get_immutable request. I said that'
[17:05] <elder> s where I was headed but you redirected me
[17:05] <elder> Just a sec, let me look at the code.
[17:06] <elder> I was sending the request to the rbd image's header (obj_md_name)
[17:06] <elder> pool "rbd" request "get_immutable_metadata"
[17:06] <elder> And the only two things I wanted back were obj_order and obj_prefix.
[17:07] <elder> He said I needed to request those individually. So I started working on just getting obj_prefix.
[17:07] <elder> It was fairly quick to throw that together. Problem at the end of Friday was that I needed to request it without having to map the image.
[17:07] <sage> elder: ah
[17:07] <elder> (BAsed on how I was settnig it up)
[17:08] <sage> elder: that's what cls_rbd_client.cc's get_immutable_metadata() does.. it builds an osd request that calls get_size and get_object_prefix
[17:08] <elder> He said mapping the image was going to be a lot of work for the new format, so he recommended I try to just do a request more directly.
[17:08] <elder> Right, and I had the impression I was somehow connecting to that interface.
[17:09] <sage> ah. that's just a client-side convenience function for the userland code
[17:09] <elder> What I had previously planned to do was somehow implement a mirror of that interface, making the get_size and get_object_prefix calls to populate it.
[17:09] <elder> But you saw my big list of functions and said that's not how I wanted to do it. Maybe you misunderstood my intention.
[17:10] <elder> The good news is I have my code talking successfully to the other end and getting a response, with no crashes at all.
[17:10] <sage> that's possible. what i thought you were doing was methods that would build entire requests for each call, though, instead of a single osd_op entry for the request's ops vector
[17:10] <elder> Oh, well I suppose that's possible but I wouldn't do that to begin with...
[17:11] <elder> It's not hard to go back to the many little requests rather than just get_immutable and get_mutable
[17:11] <elder> I just need to know what interface to program to.
[17:11] <sage> the per-call helpers should just populate an osd_op op struct and pack in the arguments into a growing data blob. and probably have a reply helper that unpacks the results
[17:12] <elder> See if I get this...
[17:13] <elder> So on the (kernel) client I will fire off multiple RPC-like requests to the server. Each will be somethng like "get size" and a separate one, "get object prefix". I aggregate those results in order to satisfy whatever requested the info.
[17:13] <elder> And of course, I have to decode each result.
[17:13] * verwilst (~verwilst@d5152FEFB.static.telenet.be) Quit (Quit: Ex-Chat)
[17:13] <elder> Or are you suggesting a single request to the server will embed requests for multiple bits of info?
[17:16] <elder> I already have code to implement just the "get object prefix" request, but it's built on infrastructure that assumes I've got an rbd_dev structure built up.
[17:17] <elder> And I believe that gets done by "mapping" the rbd image to Linux device.
[17:18] <elder> I need to do that anyway, right? If that's the case, I can just get going on that part of it, though we should probably talk about it before I go very far.
[17:18] <elder> I have not pushed my branch yet--and can if you want to see what I've got.
[17:20] <fghaas> [not trying to interrupt here, sage/elder please feel free to ignore me] when using the Python bindings, do I assume correctly that rados.Rados(rados_id='foo') would cause the client library to try to authenticate as client.foo, using whatever key is available for [client.foo]?
[17:21] <sage> elder: a single request that has two osd_ops, one for get_size and one for get_object_prefix
[17:23] <sage> fghaas: that sounds right, although i'm not very familiar with the python bindings.. joshd or dmick are the ones to ask
[17:23] <sage> they'll be in shortly
[17:23] <elder> Got it.
[17:23] <elder> sage that will make use of the ability to send multiple ops, which to date only ever uses oned.
[17:23] <elder> one
[17:23] <fghaas> sage: ok, if that's what it is, then I've either misunderstood capabilities or I'm able to do something in python that I'm not supposed to be able to do :)
[17:24] <sage> fghaas: it authenticates as client.admin by default, so if that key is available that might explain it?
[17:25] <fghaas> well I explicitly set rados_id="test" because I want to connect as client.test
[17:25] <elder> sage, in order to do that I need to do the code to map the rbd though, right? I.e., implement "echo ... > /sys/bus/rbd/add" for new-format images?
[17:26] <elder> (Well, that is, I will ultimately need to do that. I could do test code that avoids it if I wanted to.) Right?
[17:26] <sage> i can't remember how the rbd stuff is structured exactly, but this should be part of the map process. the rbd_dev can't finish initializing until you have this information
[17:26] <elder> OK.
[17:27] <elder> I will start looking at it again.
[17:27] <sage> the map/init process will: resolve name-> id, call those methods on the id object, then call the mutable methods, then finish. something like that.
[17:27] <sage> s/id object/header object/
[17:27] <elder> OK.
[17:27] <elder> Yes, I understood that.
[17:35] * rino (~rino@64.74.214.126) Quit (Quit: BitchX-1.2c01-svn -- just do it.)
[17:41] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[17:52] * laks (~laks@117.192.142.28) has joined #ceph
[17:55] * gregaf (~Adium@aon.hq.newdream.net) Quit (Quit: Leaving.)
[17:55] * Ryan_Lane (~Adium@216.38.130.162) has joined #ceph
[17:56] <laks> I was trying to test ceph,I download 0.47.3 tar file and compiling it on Fedora-16.It fails with message like
[17:56] <laks> /usr/bin/ld: ./.libs/libglobal.a(libcommon_la-HeartbeatMap.o): undefined reference to symbol 'pthread_rwlock_wrlock@@GLIBC_2.2.5'
[17:56] <laks> /usr/bin/ld: note: 'pthread_rwlock_wrlock@@GLIBC_2.2.5' is defined in DSO /lib64/libpthread.so.0 so try adding it to the linker command line
[17:57] <laks> /lib64/libpthread.so.0: could not read symbols: Invalid operation
[17:57] <laks> any thoughts on whats creating this issue? thanks for help.
[17:58] * gregaf (~Adium@aon.hq.newdream.net) has joined #ceph
[18:04] * Tv_ (~tv@aon.hq.newdream.net) has joined #ceph
[18:07] * nhm (~nh@65-128-158-48.mpls.qwest.net) has joined #ceph
[18:14] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[18:18] * sagewk (~sage@aon.hq.newdream.net) has joined #ceph
[18:26] <laks> complete error message can be found here http://pastebin.com/7cMiFHv8
[18:28] <gregaf> laks: what's that from?
[18:29] <laks> its from compiling 0.47.3 on fedora-16 machine
[18:30] * fghaas (~florian@91.119.51.139) Quit (Read error: No route to host)
[18:31] <gregaf> hrm??? that looks to me like your permissions are busted
[18:32] <gregaf> "/lib64/libpthread.so.0: could not read symbols: Invalid operation"
[18:32] <gregaf> unless that error means something else to somebody who's done more admin than I have
[18:33] <laks> hmm...I compiled it as root user..
[18:33] <laks> output from ls says "lrwxrwxrwx. 1 root root 21 Mar 2 11:00 /lib64/libpthread.so.0 -> libpthread-2.14.90.so"
[18:33] <gregaf> okay
[18:34] <gregaf> well, google says http://lists.fedoraproject.org/pipermail/devel/2010-March/133601.html
[18:36] <laks> okay,thanks seems like some issue with F16. I'll go through the link and check again.
[18:36] <gregaf> yep! :)
[18:37] <laks> thanks for the help :)
[18:38] * laks (~laks@117.192.142.28) has left #ceph
[18:38] <gregaf> np, we aim to please ;)
[18:40] * nhorman (~nhorman@hmsreliant.think-freely.org) Quit (Ping timeout: 480 seconds)
[18:52] * ninkotech (~duplo@89.177.137.231) Quit (Ping timeout: 480 seconds)
[18:55] * loicd (~loic@83.167.43.235) Quit (Quit: Leaving.)
[18:59] * nhorman (~nhorman@hmsreliant.think-freely.org) has joined #ceph
[19:01] * chutzpah (~chutz@100.42.98.5) has joined #ceph
[19:03] * ninkotech (~duplo@89.177.137.231) has joined #ceph
[19:08] * ninkotech (~duplo@89.177.137.231) Quit (Remote host closed the connection)
[19:14] * joshd (~joshd@aon.hq.newdream.net) has joined #ceph
[19:16] * ninkotech (~duplo@89.177.137.231) has joined #ceph
[19:23] * loicd (~loic@178.20.50.225) has joined #ceph
[19:38] <elder> sage, here's what I was referring to with respect to rearranging some stuff in rbd for the new format.
[19:38] <elder> Looking at the kernel drivers/block/rbd.c... rbd_add() calls rbd_get_client(), which parses options, creates a client, and so on.
[19:38] <elder> In order to look up the device id given its name, I need to read the contents of the rbd_id.<image_name> object (if it exists). So I need to send a request to an OSD to do that.
[19:39] <elder> What I'm talking about is that the code right now--in rbd_do_request(), is oriented around an existing rbd_client already having an osd client.
[19:39] <elder> I'd like to have it so I can use an existing osd client (or create a new one) but *not* have the implicit rbd_dev already initialized.
[19:39] <elder> I think it mostly means changing rbd_do_request() so that it gets passed an osd client rather than an rbd_dev, or something along those lines.
[19:39] <elder> There's more to it than that, but that's the general idea.
[19:41] <elder> sage, I'm going to go get some lunch. Back shortly.
[19:52] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[20:03] * mtk (~mtk@ool-44c35bb4.dyn.optonline.net) Quit (Remote host closed the connection)
[20:04] <sagewk> elder: i'm not sure why the existing structure is a problem..
[20:04] * MarkDude (~MT@c-71-198-138-155.hsd1.ca.comcast.net) has joined #ceph
[20:04] <sagewk> doing the name->id translation is just part of the rbd_dev init sequence, it's just growing another osd round trip in the process
[20:06] <sagewk> tv_: i was going to tag 0.48 as 0.48argonaut. so the deb will be 0.48argonaut-1, and the .1 can either be 0.48.1argonaut-1 or 0.48argonaut.1-1
[20:07] <Tv_> sagewk: no dash in that name please
[20:07] <sagewk> i want to get the codename in there somewhere so people know what they have with dpkg -l etc, but still make the versions sort
[20:07] <Tv_> sagewk: oh rigth sorry misread
[20:07] <Tv_> i see what you're saying; argonaut vs 1 order
[20:07] <sagewk> does that seem liek the right approach?
[20:08] <joao> have we settled for a code name theme?
[20:08] <sagewk> either s/48/48argonaut/, or s/-1/argonaut-1/. or maybe even append argonaut after the -debianversion..
[20:08] <Tv_> sagewk: dpkg --compare-versions is always the ultimate answer
[20:08] <sagewk> but that is usually for build/release crap
[20:09] <dmick> joao: yes
[20:09] <Tv_> sagewk: i don't really have much of an opinion on the version number, apart from technicalities like strict increasing, automatically comparable.. to me, argonaut == 0.48, having both is redundant
[20:09] * The_Bishop (~bishop@2a01:198:2ee:0:b44f:96e3:4d1f:3247) Quit (Quit: Wer zum Teufel ist dieser Peer? Wenn ich den erwische dann werde ich ihm mal die Verbindung resetten!)
[20:10] <sagewk> both 0.48argonaut.1-1 and 0.48.1argonaut-1 work sort properly.
[20:11] <Tv_> sagewk: also, always compare with '<<' due to historical baggage
[20:11] <sagewk> yeah, it's just a matter of keeping argonaut sorting wrt the existing versions.
[20:11] <Tv_> or >> etc
[20:11] <Tv_> no < no lt
[20:12] <Tv_> sagewk: i guess 0.48argonaut.1 makes it clearer that 0.48==argonaut
[20:12] <Tv_> as opposed to 0.48.1==argonaut
[20:12] <joao> dmick, oh yeah... somehow I missed Yehuda's email about that :x
[20:12] * dmick scratches head wondering what tool uses << for comparison
[20:12] <Tv_> dmick: dpkg
[20:13] <Tv_> dmick: because '<' got defined in a fubar way
[20:13] <dmick> ew
[20:14] <Tv_> "The deprecated forms < and > were used to mean earlier/later or equal, rather than strictly earlier/later, so they should not appear in new packages (though dpkg still supports them)."
[20:14] <Tv_> http://www.debian.org/doc/debian-policy/ch-relationships.html
[20:16] <sagewk> k thanks. well, either way, it'll be 0.48argonaut-1 for this release :)
[20:16] <elder> sage, I was just wanting to reuse the code that talks with the OSD's. I'm just starting to look at it so I probably have some misconceptions. But I think I'll at least encapsulate the some of the stuff inside rbd_do_request() so there remains only one place the client calls out to an OSD server. I can then use that inside rbd_add().
[20:17] <elder> 0.48argonaut-1 is a little ishy.
[20:18] <Tv_> it's a little fishy ;)
[20:19] <dmick> ishy?
[20:19] <elder> Maybe that too. The mix of alpha and numeric and the presence and absence of sort of separator characters like . and -
[20:20] <elder> But hey, not my bailiwick and that's fine by me.
[20:46] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[20:48] * adjohn (~adjohn@69.170.166.146) has joined #ceph
[21:03] * The_Bishop (~bishop@p4FCDEF6F.dip.t-dialin.net) has joined #ceph
[21:07] * fghaas (~florian@91.119.51.139) has joined #ceph
[21:13] * adjohn (~adjohn@69.170.166.146) Quit (Quit: adjohn)
[21:16] * The_Bishop (~bishop@p4FCDEF6F.dip.t-dialin.net) Quit (Ping timeout: 480 seconds)
[21:25] * The_Bishop (~bishop@p4FCDF815.dip.t-dialin.net) has joined #ceph
[21:32] * elder (~elder@c-71-195-31-37.hsd1.mn.comcast.net) Quit (Ping timeout: 480 seconds)
[21:36] * fghaas (~florian@91.119.51.139) Quit (Ping timeout: 480 seconds)
[21:49] * Dr_O (~owen@host-78-145-29-241.as13285.net) has joined #ceph
[21:50] <sagewk> joshd: few small comments on wip-rbd-id, otherwise looks good!
[21:50] <joshd> cool, thanks
[22:07] <nhm> man it's hot today
[22:15] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[22:15] * nhorman (~nhorman@hmsreliant.think-freely.org) Quit (Quit: Leaving)
[22:17] * fghaas (~florian@91.119.51.139) has joined #ceph
[22:42] * brambles (brambles@79.133.200.49) Quit (Remote host closed the connection)
[22:43] * brambles (brambles@79.133.200.49) has joined #ceph
[22:44] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[22:56] * MarkDude (~MT@c-71-198-138-155.hsd1.ca.comcast.net) Quit (Read error: Connection reset by peer)
[23:01] * fghaas (~florian@91.119.51.139) Quit (Ping timeout: 480 seconds)
[23:17] * The_Bishop (~bishop@p4FCDF815.dip.t-dialin.net) Quit (Quit: Wer zum Teufel ist dieser Peer? Wenn ich den erwische dann werde ich ihm mal die Verbindung resetten!)
[23:23] * Ryan_Lane1 (~Adium@216.38.130.162) has joined #ceph
[23:23] * Ryan_Lane (~Adium@216.38.130.162) Quit (Read error: Connection reset by peer)
[23:24] * loicd (~loic@178.20.50.225) Quit (Ping timeout: 480 seconds)
[23:26] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[23:30] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[23:33] * The_Bishop (~bishop@p4FCDF815.dip.t-dialin.net) has joined #ceph
[23:49] * adjohn (~adjohn@69.170.166.146) has joined #ceph
[23:57] * The_Bishop (~bishop@p4FCDF815.dip.t-dialin.net) Quit (Ping timeout: 480 seconds)

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