#ceph IRC Log

Index

IRC Log for 2012-07-03

Timestamps are in GMT/BST.

[0:02] * The_Bishop (~bishop@p4FCDF815.dip.t-dialin.net) has joined #ceph
[0:12] * ThoughtCoder (~ThoughtCo@202-173-147-27.mach.com.au) has joined #ceph
[0:20] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) has joined #ceph
[0:21] * loicd (~loic@magenta.dachary.org) has joined #ceph
[0:30] * sagewk (~sage@aon.hq.newdream.net) Quit (Quit: Leaving.)
[0:36] * elder (~elder@c-71-195-31-37.hsd1.mn.comcast.net) has joined #ceph
[0:40] <dmick> AGH BROWSER ACCESS TO PLANA IPMI SO SLOOOOOOOOOW
[0:41] <elder> MUST... CONTROL.... URGE.... TO....
[0:41] <dmick> KHAAAAAAAAAAAAAN!
[0:41] <elder> (It's echoing)
[0:42] <dmick> seriously. I'm in BIOS setup and every keystroke is taking about 15s to have any effect. As if this wasn't hell enough already.
[0:42] * Ryan_Lane (~Adium@216.38.130.162) has joined #ceph
[0:44] * The_Bishop shivers
[0:47] * Ryan_Lane1 (~Adium@216.38.130.162) Quit (Ping timeout: 480 seconds)
[0:48] * sagewk (~sage@aon.hq.newdream.net) has joined #ceph
[0:49] <dmick> awesome! Java is using 214% of my cpu
[0:50] <elder> I don't know if you know it, but that's a LOT.
[0:50] <lurbs> date -s "`date`"
[0:50] <lurbs> Should fix it.
[0:50] * lurbs blames leap seconds.
[0:50] <dmick> orly
[0:50] <dmick> hm
[0:50] <elder> Long day the othrday.
[0:50] <lurbs> Might also need to kick over your ntp daemon.
[0:51] <dmick> that did, in fact, make java's CPU usage drop immediately
[0:51] <dmick> who says random whining isn't productive?
[0:51] <elder> joshd, I would like to recommend we come up with a better way of specifying the "new" rbd image format. I.e., don't say "--new-format" because, well, it won't be new for very long.
[0:52] <elder> Maybe "format 2", like maybe "--format=2" or something, somehow coming up with a name/encoding for the existing one (call it "format 1" or maybe "format 0" if you're highly computer-sciencey)
[0:53] <joshd> that's just for testing purposes right now, it should definitely change
[0:53] <dmick> --format-the-second
[0:53] <elder> I'm also looking for a succinct naming convention for symbols to distinguish old from new, so could use a decision on it now.
[0:54] <elder> Not really urgent, but if we can agree on something soon it would be helpful.
[0:56] <joshd> librbd just started using old_format, making it int format = 1 for old, and 2 for new sounds possibly more clear
[0:57] <dmick> I'm gonna have to reboot. This is just not workign
[0:58] <elder> Can we agree then to use numeric format versions, as you just described? 0 can be reserved for invalid, 1 is the original (current/old) and 2 is the next one (new)?
[0:58] <joshd> for the user interface, it might make more sense to have individual features enableable, i.e. --enable-layering/--disable-layering, which both trigger the new format, but set/unset the relevant feature bit
[0:58] <elder> That's true, but only the new format would support the feature.
[0:58] <elder> format 2 or above
[0:59] <joshd> right
[0:59] <lurbs> Layering is for clones, etc?
[0:59] <elder> Yes
[0:59] <joshd> yeah, numeric format sounds good for now
[0:59] <elder> I'm going to proceed on that assumption then. "new" format is format 2.
[0:59] <joshd> ok
[1:00] <elder> Maybe we'll want a macro and if I find a need I'll make one for common use. Otherwise I'll just use 2 in places if it makes sense.
[1:00] <elder> THanks.
[1:02] <nhm> grr, firefox keeps crashing
[1:02] <nhm> I wonder if I have some bad memory or if it's just the heat.
[1:02] <elder> Damned leap seconds.
[1:02] <nhm> that too!
[1:03] <elder> Are you outside?
[1:03] * dmick (~dmick@aon.hq.newdream.net) Quit (Read error: Connection reset by peer)
[1:03] <nhm> I blame the leap second for this gtk dependency in perf too. seriously?
[1:03] * dmick (~dmick@aon.hq.newdream.net) has joined #ceph
[1:03] <nhm> elder: no, but upstairs in the converted attic of our house. The AC doesn't get up here well.
[1:03] * dmick (~dmick@aon.hq.newdream.net) Quit ()
[1:03] <elder> Good thing you get to work in your underwear.
[1:03] <nhm> it's about 88F up here now.
[1:04] <elder> Drink plenty of liquids.
[1:04] <nhm> elder: yeah, I've been hitting up the water a lot today.
[1:04] <elder> Still expecting 100 degrees Thursday?
[1:05] <elder> Oh, looks like they've downgraded that. 97.
[1:05] <nhm> elder: last I saw it was 98, but it got hotter today than they were expecting.
[1:05] <nhm> elder: in some parts of the cities it was over 100 according to the map on kare11.
[1:07] <lurbs> You crazy kids and your Fahrenheit.
[1:08] <elder> I was going to escape the heat tomorrow and help the cleanup in Cloquet or Duluth.
[1:08] <elder> OK, 37 degrees.
[1:09] <elder> Or 310.15 K
[1:11] <rturk> nice!
[1:12] <rturk> that's about 3,500 in scoville
[1:12] <elder> That's all?
[1:14] * rturk shrugs
[1:14] <rturk> that's what google said
[1:14] <elder> Their units converter failed you huh?
[1:14] <elder> Maybe Wolfram Alpha knows.
[1:17] <elder> Nope, but it is about 1 degree above optimal temperature for eating halavah.
[1:18] <elder> So you've got that, nhm. Get your halavah out and cool it down a bit. Optimal.
[1:18] <rturk> :D
[1:18] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has left #ceph
[1:25] * dmick (~dmick@aon.hq.newdream.net) has joined #ceph
[1:43] * Qten (~qgrasso@ip-121-0-1-110.static.dsl.onqcomms.net) has joined #ceph
[1:43] * Qten (~qgrasso@ip-121-0-1-110.static.dsl.onqcomms.net) Quit ()
[1:43] * Qten (~qgrasso@ip-121-0-1-110.static.dsl.onqcomms.net) has joined #ceph
[2:00] <elder> joshd, do you know why I'd get 5 bytes back from reading "rbd_id.image"?
[2:04] <elder> Nevermind. I know why.
[2:04] <elder> It's a string--4 byte length followed by two characters.
[2:04] <elder> (well, it was a one-character string last time)
[2:05] <elder> I have now successfully ascertained the rbd image id given its name using a format 2 image.
[2:05] <elder> \o/
[2:05] <elder> |
[2:06] <elder> \/
[2:06] <elder> ?
[2:06] <joshd> yay!
[2:06] <elder> And on to the next step...
[2:06] * Tv_ (~tv@aon.hq.newdream.net) Quit (Quit: Tv_)
[2:11] * `gregorg` (~Greg@78.155.152.6) has joined #ceph
[2:12] * gregorg_taf (~Greg@78.155.152.6) Quit (Read error: Connection reset by peer)
[2:22] <elder> joshd, is that rbd id a hex value, or decimal, or what?
[2:22] <elder> I think I may have seen a 'b' suggesting it's hex. But there's no other indication.
[2:23] <joshd> elder: it's a string
[2:23] <elder> Non-numeric?
[2:23] <joshd> right now it's a hex value iirc, but it may become a uuid soon
[2:23] <elder> (I know it's a string, but how should I interpret the string?)
[2:23] <elder> OK.
[2:23] <elder> So I should treat it as a string, not a number.
[2:23] <joshd> yeah
[2:23] <elder> Got it.
[2:26] * 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!)
[2:38] * yoshi (~yoshi@p22043-ipngn1701marunouchi.tokyo.ocn.ne.jp) has joined #ceph
[2:43] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[2:46] * loicd (~loic@magenta.dachary.org) has joined #ceph
[3:00] * Ryan_Lane (~Adium@216.38.130.162) Quit (Quit: Leaving.)
[3:47] * adjohn (~adjohn@69.170.166.146) Quit (Quit: adjohn)
[3:56] * ChanServ sets mode +o sage
[3:56] * sage changes topic to 'ceph development, discussion'
[3:56] * joshd (~joshd@aon.hq.newdream.net) Quit (Quit: Leaving.)
[4:02] * chutzpah (~chutz@100.42.98.5) Quit (Quit: Leaving)
[4:20] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[4:22] * loicd (~loic@magenta.dachary.org) has joined #ceph
[4:28] <elder> sage, if you're around, I'm interested to know what the best kernel branch you've been working with lately is.
[4:31] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) has joined #ceph
[4:35] * Ryan_Lane (~Adium@c-98-210-205-93.hsd1.ca.comcast.net) has joined #ceph
[5:05] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) Quit (Quit: adjohn)
[5:25] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[5:28] * loicd (~loic@magenta.dachary.org) has joined #ceph
[5:40] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) has joined #ceph
[5:41] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) Quit ()
[5:43] * mwhudson (~mwh@120.136.5.22) has joined #ceph
[5:47] * renzhi (~renzhi@208.53.156.36) Quit (Ping timeout: 480 seconds)
[5:59] <sage> elder: testing!
[5:59] <elder> OK.
[6:00] <elder> Turns out it doesn't matter. I'm hitting a problem that's crashing my UML and I thought it might be because of a closed connection. EIther way it's a problem, but I was hoping to avoid any spurious problems.
[6:00] <elder> (I.e. not directly related to what I'm working on)
[6:05] <ThoughtCoder> I'm seeing: HEALTH_WARN 1 pgs stale; 1 pgs stuck stale - after replacing a faulty/flapping OSD. The stale PG is on the, now functioning, osd (osd.2) - Is there a way to force this PG to active?
[6:07] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[6:08] <dmick> what does ceph -s say about the OSD?
[6:09] * loicd (~loic@magenta.dachary.org) has joined #ceph
[6:13] <dmick> ThoughtCoder: ^
[6:14] <ThoughtCoder> Sorry - will look!
[6:15] <ThoughtCoder> osd e1737: 23 osds: 23 up, 23 in
[6:16] <ThoughtCoder> Any other tips?
[6:17] <dmick> "ceph pg <num> query" might give some clues
[6:20] <ThoughtCoder> Sorry for the stupidity, how to I obtain the <num> value?
[6:21] <dmick> That's the PG <num>. It should be mentioned in ceph health (maybe you need ceph health detail)
[6:21] <ThoughtCoder> Cheers - I get this: pg 3.5 is stuck stale+active+clean, last acting [2]
[6:21] <ThoughtCoder> Running the query now
[6:22] <ThoughtCoder> But doesn't seem very happy
[6:23] <dmick> mostly interested in recovery_state, I think
[6:24] <dmick> what does it say?
[6:25] <ThoughtCoder> It's not even printing out
[6:26] <dmick> ceph pg 3.5 query hung?
[6:27] <dmick> well
[6:27] <dmick> I have to go
[6:27] <dmick> good luck. there's some helpful info at http://ceph.com/docs/master/ops/manage/failures/osd/
[6:28] <ThoughtCoder> Thanks
[6:28] * dmick (~dmick@aon.hq.newdream.net) Quit (Quit: Leaving.)
[6:32] <sage> elder: how are you getting it to crash?
[6:33] <sage> stuck stale implies that the osd(s) storing that pg are *all* down, hence pg .. query hanging
[6:33] <sage> thoughtcoder: ^
[6:34] <sage> thoughtcoder: ceph pg map 3.5 to see which osds should be responsible, and then make sure they're running. maybe restart them if they are.
[6:42] * sage (~sage@cpe-76-94-40-34.socal.res.rr.com) Quit (Remote host closed the connection)
[6:45] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) has joined #ceph
[6:56] * alexxy (~alexxy@2001:470:1f14:106::2) Quit (Ping timeout: 480 seconds)
[6:57] * adjohn (~adjohn@50-0-133-101.dsl.static.sonic.net) Quit (Quit: adjohn)
[7:01] * alexxy (~alexxy@2001:470:1f14:106::2) has joined #ceph
[7:17] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[7:20] * loicd (~loic@magenta.dachary.org) has joined #ceph
[7:22] * chuanyu (chuanyu@linux3.cs.nctu.edu.tw) has joined #ceph
[7:23] * mwhudson (~mwh@120.136.5.22) Quit (Ping timeout: 480 seconds)
[8:17] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[8:21] * Ryan_Lane (~Adium@c-98-210-205-93.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[8:38] <ThoughtCoder> @sage: I can't even get ceph pg map 3.5 tells me it's on osd.2 - does this mean that osd.2 is completely dead?
[8:38] <ThoughtCoder> *sorry, delete everything before pg map...
[8:39] <ThoughtCoder> Bear in mind I had issue with this OSD earlier in the day.
[9:10] * s[X] (~sX]@eth589.qld.adsl.internode.on.net) Quit (Remote host closed the connection)
[9:13] * pmjdebruijn (~pmjdebrui@overlord.pcode.nl) has joined #ceph
[9:13] <pmjdebruijn> hi guys
[9:13] <pmjdebruijn> any particular reason why the new version is 0.48argonaut?
[9:13] <pmjdebruijn> what will 0.48.1 be called?
[9:13] <pmjdebruijn> 0.48.1argonaut?
[9:13] <pmjdebruijn> 0.48argonaut.1 ?
[9:13] <pmjdebruijn> :)
[9:14] <pmjdebruijn> I mean I get the code name thing, but I'm wondering why that has to be part of the version number
[9:17] * loicd (~loic@83.167.43.235) has joined #ceph
[9:24] * fghaas (~florian@91.119.51.139) has joined #ceph
[9:29] * ThoughtCoder (~ThoughtCo@202-173-147-27.mach.com.au) Quit (Ping timeout: 480 seconds)
[9:45] * verwilst (~verwilst@66.14-246-81.adsl-static.isp.belgacom.be) has joined #ceph
[9:50] * fghaas (~florian@91.119.51.139) Quit (Ping timeout: 480 seconds)
[10:10] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[10:13] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[10:17] * fghaas (~florian@91.119.51.139) has joined #ceph
[10:40] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[10:42] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[10:44] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[10:48] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[10:57] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[11:05] * yoshi (~yoshi@p22043-ipngn1701marunouchi.tokyo.ocn.ne.jp) Quit (Remote host closed the connection)
[11:10] * morse (~morse@supercomputing.univpm.it) Quit (Quit: Bye, see you soon)
[11:10] * morse_ (~morse@supercomputing.univpm.it) has joined #ceph
[11:18] * morse_ is now known as morse
[11:20] * lofejndif (~lsqavnbok@9KCAAGGTE.tor-irc.dnsbl.oftc.net) has joined #ceph
[11:22] * brambles (brambles@79.133.200.49) Quit (Quit: leaving)
[11:23] * brambles (brambles@79.133.200.49) has joined #ceph
[12:42] * ThoughtCoder (~ThoughtCo@60-240-78-43.static.tpgi.com.au) has joined #ceph
[12:52] * nhorman (~nhorman@2001:470:8:a08:7aac:c0ff:fec2:933b) has joined #ceph
[13:37] * loicd (~loic@83.167.43.235) Quit (Ping timeout: 480 seconds)
[13:42] * fghaas (~florian@91.119.51.139) Quit (Ping timeout: 480 seconds)
[13:52] * ThoughtCoder (~ThoughtCo@60-240-78-43.static.tpgi.com.au) Quit (Ping timeout: 480 seconds)
[14:18] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[14:20] * alexxy (~alexxy@2001:470:1f14:106::2) Quit (Ping timeout: 480 seconds)
[14:20] * alexxy (~alexxy@masq119.gtn.ru) has joined #ceph
[14:24] * alexxy[home] (~alexxy@2001:470:1f14:106::2) has joined #ceph
[14:29] * alexxy (~alexxy@masq119.gtn.ru) Quit (Ping timeout: 480 seconds)
[14:41] * Kioob`Taff1 (~plug-oliv@local.plusdinfo.com) has joined #ceph
[14:59] * loicd (~loic@83.167.43.235) has joined #ceph
[15:02] <elder> sagewk, in answer to your question above, whenever you get in...
[15:06] <elder> I'm experimenting with calling the get_size and get_object_prefix methods on the header object.
[15:07] <elder> I can see I got EINVAL back on the first one, not sure beyond that, but it ends up sort of sitting there for a few seconds, then panicking, maybe due to a timeout or something.
[15:07] <elder> I'll rebase my work on the testing branch, just in case.
[15:13] <elder> I looked at (but wouldn't call it a review) your few changes on the top of testing, sagewk. They make good sense. I know I had caused that breakage in my attempt to consolidate the initialization of a connection. I was doing it to refactor, but wasn't thinking about reconnects being a reopen.
[15:14] <elder> In hindsight I realize that an open operation is really just setting the peer address and allowing a connection to get established. The real work happens on the first send.
[15:16] * mtk (~mtk@ool-44c35bb4.dyn.optonline.net) has joined #ceph
[15:40] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[15:42] * fghaas (~florian@91.119.51.139) has joined #ceph
[16:13] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[16:27] * lofejndif (~lsqavnbok@9KCAAGGTE.tor-irc.dnsbl.oftc.net) Quit (Quit: gone)
[17:13] * verwilst (~verwilst@66.14-246-81.adsl-static.isp.belgacom.be) Quit (Quit: Ex-Chat)
[17:13] <elder> sagewk, josh, is it possible that a method call writes the data it returns somewhere *other* than in the data portion of the message that gets sent back?
[17:15] <elder> Do I need to supply a reply message buffer of some kind? I have been successful reading the rbd_id object's content using the message data. The rbd_header object has 0 bytes of content. But I'm getting errors when trying to read data back (the same way I do for a read request) when doing a method call, like "get_size".
[17:23] * bchrisman (~Adium@c-76-103-130-94.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[17:25] * sage_ (~sage@2600:1012:b00f:684b:1cf1:c8e2:e7da:3028) has joined #ceph
[17:27] * Tv_ (~tv@aon.hq.newdream.net) has joined #ceph
[17:28] * fghaas (~florian@91.119.51.139) Quit (Quit: Leaving.)
[17:40] * sage_ (~sage@2600:1012:b00f:684b:1cf1:c8e2:e7da:3028) Quit (Ping timeout: 480 seconds)
[17:42] <sagewk> elder: hmm, yeah the reply gets preallocated i think. looking
[17:44] <sagewk> elder: yeah, you need to populate req->r_pages, r_num_pages, r_page_alignment with space for the reply
[17:45] <sagewk> see osd_client.c get_reply for where this is handed to the msgr, and ceph_osdc_alloc_request
[17:45] <elder> I've got that.
[17:45] <sagewk> tho it looks like the latter only set's r_pages for you, you have to populate the other fields
[17:45] <elder> I'm passing a page (or maybe multiple).
[17:46] <elder> Wait, let me look.
[17:48] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[17:48] <elder> I'm setting r_pages. I don't think I'm setting r_num_pages. That function should be setting it at the same time.
[17:48] <elder> (A mon avis)
[17:48] <elder> Sorry, you don't speak forn.
[17:48] <elder> In my opinion.
[17:49] <elder> Anyway, that could be it. I'll look a little closer when I'm ready to try it again.
[17:49] <elder> Rearranging stuff in my patch series at the moment.
[17:55] * loicd (~loic@83.167.43.235) Quit (Quit: Leaving.)
[17:59] * darkfader (~floh@188.40.175.2) Quit (Remote host closed the connection)
[17:59] * guilhemfr (~guilhem@AMontsouris-652-1-50-250.w92-163.abo.wanadoo.fr) has joined #ceph
[17:59] * darkfader (~floh@188.40.175.2) has joined #ceph
[18:00] <guilhemfr> hi all
[18:00] <guilhemfr> I have some pb with ceph conf
[18:01] <guilhemfr> I can't use mon addr with a dns record
[18:02] <guilhemfr> unable to parse address for mon.d: addr='ceph-mon1:6789'
[18:02] <guilhemfr> "mon addr = ceph-mon1:6789"
[18:04] <guilhemfr> ceph-fuse seems to work with dns name : http://tracker.newdream.net/issues/321
[18:05] <guilhemfr> but here I try with ceph-osd
[18:06] <joao> the parsing function assumes it is an ipv4 or ipv6 address, afaict
[18:06] <guilhemfr> it's VERY limited in my case
[18:07] <guilhemfr> all my VM have is addr by dhcp
[18:07] <guilhemfr> and update a dynamic dns
[18:08] <joao> I'd suggest opening a feature request for it (not sure if one is already open or not)
[18:08] <guilhemfr> (and a VM can't keep its addr because routing is done by subnet between host)
[18:09] <Tv_> guilhemfr: ceph really won't work well with dynamic addresses; it is built for data center networks
[18:09] <guilhemfr> it's a data center network
[18:10] <guilhemfr> but I want to use VM to do live migration
[18:10] <Tv_> guilhemfr: umm, you probably don't want to run ceph inside virtualization
[18:10] <Tv_> the disk io will be miserable
[18:10] <Tv_> ceph is the thing that *provides* live migration capabilities
[18:11] <guilhemfr> virtio + io=native
[18:11] <guilhemfr> and you don't have latency on disk
[18:11] <guilhemfr> and here it's only for mon
[18:11] <Tv_> well mon can't change it's ip address anyway
[18:11] <Tv_> the ip address is its identity
[18:11] <guilhemfr> :/
[18:11] <guilhemfr> why to have an "id" ?
[18:12] <Tv_> because everyone needs to know where the mon is, anyway
[18:12] <guilhemfr> yes
[18:12] <Tv_> dns ttl is not a viable traffic redirection mechanism
[18:12] <guilhemfr> and the mon is at $DNSNAME
[18:13] <guilhemfr> I can understand that my mon can be unreachable if I migrate it
[18:13] <guilhemfr> And must restart my cluster to understand this change
[18:13] <guilhemfr> but put an ipaddress... quite painfull
[18:14] <Tv_> there are enough problems in distributed software even without trying to solve mobile ip agents at the same time
[18:14] * dmick (~Dan@cpe-76-87-42-76.socal.res.rr.com) has joined #ceph
[18:15] <guilhemfr> so why #321 issue have resolved it for fuse ?
[18:15] <Tv_> clients are different
[18:15] <Tv_> and a lot simpler
[18:16] <Tv_> even if putting a dns name in the config had worked
[18:16] <guilhemfr> I can understand some problem with dns
[18:16] <Tv_> what do you expect to happen when the dhcp servers gives the computer a different lease?
[18:16] <guilhemfr> but you can do like nginx
[18:17] <Tv_> how do you expect everyone else to notice when you live migrate the mon?
[18:17] <guilhemfr> nginx doesn't update dns name according to ttl
[18:17] <guilhemfr> it resolve addr one time for all at startup
[18:17] <Tv_> dns lookups required for daemons to start is one pet peeve of mine
[18:17] <Tv_> that's like one of the most typical "fixes" you need to do to an apache installation
[18:18] * dmick (~Dan@cpe-76-87-42-76.socal.res.rr.com) Quit ()
[18:18] <Tv_> a single reboot when the switch is rebooting at the same time (like, after a power outage) means nothing works
[18:19] * sage_ (~sage@aon.hq.newdream.net) has joined #ceph
[18:23] * The_Bishop (~bishop@2a01:198:2ee:0:9157:ae54:8a72:ea5c) has joined #ceph
[18:24] <guilhemfr> of course it's not a solution
[18:24] <guilhemfr> a *good
[18:25] <guilhemfr> and I can understand that many people in big data center
[18:25] * Dr_O (~owen@host-78-145-29-241.as13285.net) Quit (Read error: Connection reset by peer)
[18:25] <guilhemfr> will used IP
[18:26] <guilhemfr> but, in my case, DNS is mandatory :/
[18:27] * Dr_O (~owen@host-78-145-29-241.as13285.net) has joined #ceph
[18:32] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[18:33] * Ryan_Lane (~Adium@216.38.130.162) has joined #ceph
[18:35] * bchrisman (~Adium@108.60.121.114) has joined #ceph
[18:39] * guilhemfr (~guilhem@AMontsouris-652-1-50-250.w92-163.abo.wanadoo.fr) Quit (Remote host closed the connection)
[18:52] * alexxy[home] (~alexxy@2001:470:1f14:106::2) Quit (Read error: Connection reset by peer)
[18:52] * alexxy (~alexxy@2001:470:1f14:106::2) has joined #ceph
[18:58] * yehudasa_ (~yehudasa@aon.hq.newdream.net) has joined #ceph
[19:16] * joshd (~joshd@aon.hq.newdream.net) has joined #ceph
[19:16] * chutzpah (~chutz@100.42.98.5) has joined #ceph
[19:32] * cephalobot (~ceph@ps94005.dreamhost.com) has joined #ceph
[19:34] * cephalobot (~ceph@ps94005.dreamhost.com) Quit ()
[19:34] * cephalobot (~ceph@ps94005.dreamhost.com) has joined #ceph
[19:36] <joao> btw, does anyone has any idea why do we have constants with the format CEPH_MSG_AUTH or CEPH_MSG_MON_* and MSG_MON_* ?
[19:36] <joao> s/has/have
[19:40] <gregaf> "have" is correct
[19:40] <gregaf> and I'm not sure what question you're asking
[19:40] <gregaf> you mean the inconsistency about CEPH_MSG versus MSG?
[19:42] <joao> yeah
[19:42] <sagewk> CEPH_MSG ones are shared by the kernel and are in msgr.h; the others weren't updated bc i was lazy
[19:44] <joao> so, if a patch updated them all nothing of value would be lost? (aside from the time it took)
[19:44] * loicd (~loic@magenta.dachary.org) has joined #ceph
[19:46] * ThoughtCoder (~ThoughtCo@60-240-78-43.static.tpgi.com.au) has joined #ceph
[19:47] <rturk> Everyone welcome cephalobot to the channel :)
[19:47] <rturk> not to be completely confused with cephlogbot
[19:47] <rturk> only partially confused
[19:48] <joao> lol
[19:48] <joao> does it do neat tricks?
[19:49] <rturk> I'm sure it does, but the only trick I configured was the "log everything" trick
[19:50] <joao> oh :(
[19:50] <yehudasa_> so that's really confusing now
[19:50] <rturk> there are some interesting activity analytics I'd like to run, and the cephlogbot format isn't supported by the tools I have in mind
[19:50] <rturk> yeah
[19:50] <rturk> I know
[19:50] <gregaf> we should get interactive chatting on build status and stuff :D
[19:51] <rturk> I think that would be awesome. I'd also like to enable some meeting note functionality
[19:51] <rturk> someday :)
[19:51] <nhm> rturk: offtopic chat metrics? ;)
[19:51] <joao> better yet, it should schedule teuthology runs
[19:52] <rturk> joao: that would be really cool
[19:52] <rturk> or a redmine integration
[19:52] <joao> and tell us when they were ready
[19:52] <rturk> nhm: yeah, I'm going to measure the amount of conversation that's not about Ceph
[19:52] <yehudasa_> .. or when jenkins runs failed
[19:52] <Tv_> .. or spam the channel so much i stopped reading.. ;)
[19:53] <rturk> Tv_: that can be arranged
[19:53] <elder> How will you know what's about ceph and what is not?
[19:53] <rturk> elder: I'm just kidding
[19:53] <rturk> but just in case I'm not, you should make sure to put "(ceph)" at the end of every chat line that concerns ceph.
[19:54] <joao> rofl
[19:54] <elder> OK. (not ceph)
[19:54] * ThoughtCoder (~ThoughtCo@60-240-78-43.static.tpgi.com.au) Quit (Ping timeout: 480 seconds)
[19:54] <yehudasa_> yes! (ceph)
[19:55] <Tv_> hey we could use hashtags, just like on twitter!
[19:55] <Tv_> and then, perhaps, we could organize our conversations around those hash tags
[19:55] <elder> And change all our names to @Tv_!!!
[19:55] <Tv_> and perhaps, call that a channel!
[19:55] * adjohn (~adjohn@69.170.166.146) has joined #ceph
[19:55] <Tv_> imagine that, a #ceph that would contain ceph conversation!
[19:55] <rturk> if only the server could just show us topics we subscribe to, then
[19:55] <elder> Tweet EVERYTHING!!! And add it to our facebook feed (not ceph)
[19:55] <rturk> kind of like "joining" a channel
[19:56] <elder> rturk, you have reduced the ceph content ratio. (not ceph)
[19:57] <Tv_> cephalobot: punish rturk
[19:57] <cephalobot> Tv_: Error: "punish" is not a valid command.
[19:57] <rturk> oh no, don't try to talk to it. I have no idea what its defaults allow it to do
[19:57] <Tv_> cephalobot: sudo rm -rf /
[19:57] <cephalobot> Tv_: Error: "sudo" is not a valid command.
[19:57] <rturk> lol
[20:01] <joao> rotfl
[20:01] <joao> oh well, rturk, what bot is it running?
[20:02] <rturk> supybot
[20:02] <rturk> I'm sure it does all sorts of interesting stuff
[20:02] * nhm (~nh@65-128-158-48.mpls.qwest.net) Quit (Ping timeout: 480 seconds)
[20:03] <yehudasa_> shouldn't we have a topic set for the channel now?
[20:04] <joao> I find it amusing that the channel was registered just a couple of weeks ago, and we've been using it for ages
[20:04] <rturk> yehudasa_: yeah, I think we probably should
[20:05] <yehudasa_> rturk: I think sage can get op for this channel
[20:05] * rturk notices for the first time that nobody here has chanops
[20:14] <joao> Chanserv states Sage as the one-and-only channel master
[20:20] <gregaf> wait, why doesn't he have op? he did a few days ago
[20:20] * bchrisman (~Adium@108.60.121.114) Quit (Quit: Leaving.)
[20:21] <rturk> gregaf: he did have ops a few days ago
[20:22] * nhm (~nh@65-128-158-48.mpls.qwest.net) has joined #ceph
[20:23] <nhm> ugh, I think it's time to replace this cpu or memory.
[20:25] <joao> or both, and throw a new graphics card while you're at it :p
[20:25] <nhm> joao: excellent idea
[20:25] <nhm> teehee
[20:26] <nhm> "Need a new computer for work!"
[20:27] <joao> I suspect that my bank is suffering from issues with the leap second
[20:28] <joao> their site is in maintenance, and the phone support state they've been having problems for the last couple of days
[20:35] <nhm> joao: hopefulyl you will have money left when they come back. ;)
[20:37] <joao> I already called the 24-hour service and they confirmed the money is still there
[20:38] <joao> at least that :p
[20:51] * MarkDude (~MT@c-71-198-138-155.hsd1.ca.comcast.net) has joined #ceph
[20:58] * cattelan_away is now known as cattelan
[21:03] * sagewk (~sage@aon.hq.newdream.net) Quit (Quit: Leaving)
[21:03] * cattelan is now known as cattelan_away
[21:08] * yehudasa_ (~yehudasa@aon.hq.newdream.net) Quit (Ping timeout: 480 seconds)
[21:13] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) has joined #ceph
[21:13] * sagewk (~sage@aon.hq.newdream.net) has joined #ceph
[21:19] * cattelan_away is now known as cattelan
[21:37] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[21:37] * loicd (~loic@magenta.dachary.org) has joined #ceph
[21:47] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[21:58] <elder> joshd, sage_ it just occurred to me that from the perspective of namei() lookups we might be better off naming RBD objects with the object id *first*. E.g. <object>.rbd_id, <object>.rbd_header, <object>.rbd_data.<num>
[22:00] <sagewk> elder: a fixed prefix lets us write osd/rados capabilities that restrict access in specific ways
[22:00] <sagewk> matching suffixes isn't implemented, and is more awkward
[22:00] <elder> OK, well then there's a good reason to do it the way it's now done...
[22:00] <elder> It was a small consideration anyway.
[22:01] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has left #ceph
[22:03] * begaei (~begaei@19NAAANP5.tor-irc.dnsbl.oftc.net) has joined #ceph
[22:03] * neerbeer (~Adium@65-125-22-154.dia.static.qwest.net) has joined #ceph
[22:09] <neerbeer> I've got a two host ceph setup w/ repl = 2 . This ceph SAN is mounted via another server via rbd kernel mod . I can echo 'mon_ip ' to /proc to add the rbd device just fine and I can mkfs.ext4 and all is well. I shut down the osd on the ceph host that is not running a mon to test redundancy. What is the type of behavior I should have observed given a loss of an osd.
[22:10] * darkfaded (~floh@188.40.175.2) has joined #ceph
[22:12] * alexxy (~alexxy@2001:470:1f14:106::2) Quit (Ping timeout: 480 seconds)
[22:15] * Ryan_Lane1 (~Adium@216.38.130.162) has joined #ceph
[22:15] * Ryan_Lane (~Adium@216.38.130.162) Quit (Read error: Connection reset by peer)
[22:16] * darkfader (~floh@188.40.175.2) Quit (Ping timeout: 480 seconds)
[22:20] * alexxy (~alexxy@79.173.81.171) has joined #ceph
[22:33] * cattelan is now known as cattelan_away
[22:39] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[22:40] <elder> sage_, is there an equivalent to the image id on existing rbd format? Or is it just sort of implied by the block name that's recorded in the header?
[22:40] * loicd (~loic@magenta.dachary.org) has joined #ceph
[22:43] <begaei> fuck Justin Bieber
[22:43] <begaei> Justin Bieber is a faggot
[22:43] <begaei> Justin Bieber is an animal
[22:43] <begaei> Justin Bieber is a piece of shit
[22:44] <elder> What do you think of Justin Bieber, begaei?
[22:44] <begaei> I think I already told
[22:45] <elder> So you're saying your feelings about him are on the whole, not positive?
[22:45] * begaei (~begaei@19NAAANP5.tor-irc.dnsbl.oftc.net) Quit (autokilled: This host violated network policy. If you feel an error has been made, please contact support@oftc.net, thanks. (2012-07-03 20:45:08))
[22:45] <joao> someone didn't get a ticket for Justin Bieber's show
[22:45] <joshd> elder: the latter - it's not stored explicitly in format 1
[22:45] * neerbeer (~Adium@65-125-22-154.dia.static.qwest.net) Quit (Quit: Leaving.)
[22:46] * darkfaded (~floh@188.40.175.2) Quit (Ping timeout: 480 seconds)
[22:46] <elder> OK, that's what it looked like to me but I wanted to be sure.
[22:51] * nhorman (~nhorman@2001:470:8:a08:7aac:c0ff:fec2:933b) Quit (Quit: Leaving)
[22:51] <nhm> yikes, /etc/ceph disappeared
[22:52] <sagewk> nhm: did you run chef?
[22:52] <nhm> sagewk: nope
[22:53] <sagewk> k. fwiw, the chef solo thing now removes ceph debs and clobbers /etc/ceph. there were polluted planas that were screwing up qa
[22:53] <nhm> looks like I've still got the ceph debs. Strange
[22:54] <sagewk> elder: image id is like the prefix. no direct equivalent
[22:59] * plastics (~plastics@c-69-138-42-222.hsd1.tn.comcast.net) Quit (Ping timeout: 480 seconds)
[23:00] * adjohn (~adjohn@69.170.166.146) Quit (Quit: adjohn)
[23:03] * adjohn (~adjohn@69.170.166.146) has joined #ceph
[23:03] * darkfader (~floh@188.40.175.2) has joined #ceph
[23:04] * adjohn (~adjohn@69.170.166.146) Quit ()
[23:07] * lxo (~aoliva@lxo.user.oftc.net) Quit (Ping timeout: 480 seconds)
[23:11] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[23:12] * adjohn (~adjohn@69.170.166.146) has joined #ceph
[23:16] * lxo (~aoliva@lxo.user.oftc.net) has joined #ceph
[23:18] * adjohn (~adjohn@69.170.166.146) Quit (Quit: adjohn)
[23:21] * loicd (~loic@magenta.dachary.org) Quit (Quit: Leaving.)
[23:22] * loicd (~loic@magenta.dachary.org) has joined #ceph
[23:26] * bchrisman (~Adium@c-76-103-130-94.hsd1.ca.comcast.net) has joined #ceph
[23:31] <yehudasa> sagewk / gregaf: wip-2666 could use another pair of eyeballs before merging
[23:40] <sagewk> added to my list :)
[23:48] * cattelan_away is now known as cattelan

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