Re: rbd export speed limit
Hi, Am 12.02.2013 21:45, schrieb Andrey Korolyov: > you may be interested in throttle(1) as a side solution with stdout > export option. What's throttle? Never seen this. Wouldn't it be possible to use tc? > By the way, on which interconnect you have manage to > get such speeds, Bonded Intel 2x 10GBE > if you mean 'commited' bytes(e.g. not almost empty > allocated image)? Yes commited images. Stefan -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: File exists not handled in 0.48argonaut1
I tried removing the corrupt log file (and all corrupt_log files) with no success. I then tried doing a journal flush and mkjournal. Here is the log file after trying that (with the enhanced logging turned on - ~50K lines) http://dl.dropbox.com/u/766198/ceph-2.log.gz On Mon, Feb 11, 2013 at 5:20 PM, Samuel Just wrote: > The actual problem appears to be a corrupted log file. You should > rename out of the way the directory: > /mnt/osd97/current/corrupt_log_2013-02-08_18:50_2.fa8. Then, restart > the osd with debug osd = 20, debug filestore = 20, and debug ms = 1 in > the [osd] section of the ceph.conf. > -Sam > > On Mon, Feb 11, 2013 at 2:21 PM, Mandell Degerness > wrote: >> Since the attachment didn't work, apparently, here is a link to the log: >> >> http://dl.dropbox.com/u/766198/error17.log.gz >> >> On Mon, Feb 11, 2013 at 1:42 PM, Samuel Just wrote: >>> I don't see the more complete log. >>> -Sam >>> >>> On Mon, Feb 11, 2013 at 11:12 AM, Mandell Degerness >>> wrote: Anyone have any thoughts on this??? It looks like I may have to wipe out the OSDs effected and rebuild them, but I'm afraid that may result in data loss because of the old OSD first crush map in place :(. On Fri, Feb 8, 2013 at 1:36 PM, Mandell Degerness wrote: > We ran into an error which appears very much like a bug fixed in 0.44. > > This cluster is running version: > > ceph version 0.48.1argonaut > (commit:a7ad701b9bd479f20429f19e6fea7373ca6bba7c) > > The error line is: > > Feb 8 18:50:07 192.168.8.14 ceph-osd: 2013-02-08 18:50:07.545682 > 7f40f9f08700 0 filestore(/mnt/osd97) error (17) File exists not > handled on operation 20 (11279344.0.0, or op 0, counting from 0) > > A more complete log is attached. > > First question: is this a know bug fixed in more recent versions? > > Second question: is there any hope of recovery? -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Links to various language bindings
Also, be sure to open bugs and assign them to me. On Tue, Feb 12, 2013 at 12:29 PM, Josh Durgin wrote: > On 02/08/2013 01:06 AM, Wido den Hollander wrote: >> >> Hi, >> >> I knew that there were Java bindings for RADOS, but they weren't linked. >> >> Well, some searching on Github lead me to Noah's bindings [0], but it >> was a bit of searching. >> >> I expect new users to be less fortunate and end up searching endlessly >> for them. >> >> The docs say this now: >> http://ceph.com/docs/master/api/#rados-object-store-apis >> >> Only the libcephfs bindings are linked to, but they are part of the main >> repository (which still puzzles me..). >> >> For what I know bindings exist for Java[0], PHP[1] and Python, but it's >> not to be found anywhere. >> >> Shall I submit a patch to put this in the docs or should this go on >> ceph.com itself? > > > I think it'd be good to include in the docs. > > There are also Erlang bindings written against the current api: > https://github.com/renzhi/erlrados > > There are a couple written against the older api, which wouldn't work > anymore, but wouldn't be too hard to update: > > Ruby: https://github.com/johnl/desperados > Haskell: https://github.com/athanatos/librados.hsc > > >> I'd go for the docs so we can also include some simple samples for >> people who are less experienced with Ceph/RADOS and just want to get >> started, maybe developers who's only task is just to work with RADOS. >> >> Wido >> >> [0]: https://github.com/noahdesu/java-rados >> [1]: https://github.com/ceph/phprados > > > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- John Wilkins Senior Technical Writer Intank john.wilk...@inktank.com (415) 425-9599 -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: primary pg and replica pg are placed on the same node under "ssd-primary" rule
Unfortunately this part doesn't fit into the CRUSH language. In order to do it and segregate properly by node you need to have separate SSD and HDD nodes, rather than interspersing them. (Or if you were brave you could set up some much more specified rules and pull each replica from a different rack/row, but that would take a bit of work and enough machines for that mechanism to be useful.) -Greg On Tue, Feb 12, 2013 at 12:40 PM, ymorita...@gmail.com wrote: > Hi, > > I am trying to test a storage tiering using the following > "ssd-primary" rule set introduced in the Ceph page. However, there > seems to be a chance that primary pg and replica pg are placed on the > same node. > > Is there any way to avoid this at this point in time? > > rule ssd-primary { > ruleset 4 > type replicated > min_size 0 > max_size 10 > step take ssd > step chooseleaf 1 type host > step emit > step take platter > step chooseleaf -1 type host > step emit > } > > Ceph - CRUSH Maps > http://ceph.com/docs/master/rados/operations/crush-map/#placing-different-pools-on-different-osds > > Thank you. > Yuji > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: rbd export speed limit
Hi Stefan, you may be interested in throttle(1) as a side solution with stdout export option. By the way, on which interconnect you have manage to get such speeds, if you mean 'commited' bytes(e.g. not almost empty allocated image)? On Wed, Feb 13, 2013 at 12:22 AM, Stefan Priebe wrote: > Hi, > > is there a speed limit option for rbd export? Right now i'm able to produce > several SLOW requests from IMPORTANT valid requests while just exporting a > snapshot which is not really important. > > rbd export runs with 2400MB/s and each OSD with 250MB/s so it seems to block > valid normal read / write operations. > > Greets, > Stefan > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Links to various language bindings
On 02/08/2013 01:06 AM, Wido den Hollander wrote: Hi, I knew that there were Java bindings for RADOS, but they weren't linked. Well, some searching on Github lead me to Noah's bindings [0], but it was a bit of searching. I expect new users to be less fortunate and end up searching endlessly for them. The docs say this now: http://ceph.com/docs/master/api/#rados-object-store-apis Only the libcephfs bindings are linked to, but they are part of the main repository (which still puzzles me..). For what I know bindings exist for Java[0], PHP[1] and Python, but it's not to be found anywhere. Shall I submit a patch to put this in the docs or should this go on ceph.com itself? I think it'd be good to include in the docs. There are also Erlang bindings written against the current api: https://github.com/renzhi/erlrados There are a couple written against the older api, which wouldn't work anymore, but wouldn't be too hard to update: Ruby: https://github.com/johnl/desperados Haskell: https://github.com/athanatos/librados.hsc I'd go for the docs so we can also include some simple samples for people who are less experienced with Ceph/RADOS and just want to get started, maybe developers who's only task is just to work with RADOS. Wido [0]: https://github.com/noahdesu/java-rados [1]: https://github.com/ceph/phprados -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
rbd export speed limit
Hi, is there a speed limit option for rbd export? Right now i'm able to produce several SLOW requests from IMPORTANT valid requests while just exporting a snapshot which is not really important. rbd export runs with 2400MB/s and each OSD with 250MB/s so it seems to block valid normal read / write operations. Greets, Stefan -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: slow requests, hunting for new mon
On Tue, Feb 12, 2013 at 06:28:15PM +1100, Chris Dunlop wrote: > Hi, > > What are likely causes for "slow requests" and "monclient: hunting for new > mon" messages? E.g.: > > 2013-02-12 16:27:07.318943 7f9c0bc16700 0 monclient: hunting for new mon > ... > 2013-02-12 16:27:45.892314 7f9c13c26700 0 log [WRN] : 6 slow requests, 6 > included below; oldest blocked for > 30.383883 secs > 2013-02-12 16:27:45.892323 7f9c13c26700 0 log [WRN] : slow request 30.383883 > seconds old, received at 2013-02-12 16:27:15.508374: > osd_op(client.9821.0:122242 rb.0.209f.74b0dc51.0120 [write > 921600~4096] 2.981cf6bc) v4 currently no flag points reached > 2013-02-12 16:27:45.892328 7f9c13c26700 0 log [WRN] : slow request 30.383782 > seconds old, received at 2013-02-12 16:27:15.508475: > osd_op(client.9821.0:122243 rb.0.209f.74b0dc51.0120 [write > 987136~4096] 2.981cf6bc) v4 currently no flag points reached > 2013-02-12 16:27:45.892334 7f9c13c26700 0 log [WRN] : slow request 30.383720 > seconds old, received at 2013-02-12 16:27:15.508537: > osd_op(client.9821.0:122244 rb.0.209f.74b0dc51.0120 [write > 1036288~8192] 2.981cf6bc) v4 currently no flag points reached > 2013-02-12 16:27:45.892338 7f9c13c26700 0 log [WRN] : slow request 30.383684 > seconds old, received at 2013-02-12 16:27:15.508573: > osd_op(client.9821.0:122245 rb.0.209f.74b0dc51.0122 [write > 1454080~4096] 2.fff29a9a) v4 currently no flag points reached > 2013-02-12 16:27:45.892341 7f9c13c26700 0 log [WRN] : slow request 30.328986 > seconds old, received at 2013-02-12 16:27:15.563271: > osd_op(client.9821.0:122246 rb.0.209f.74b0dc51.0122 [write > 1482752~4096] 2.fff29a9a) v4 currently no flag points reached Sorry, I forgot to add what lead me to look into this in the first place... The b2 machine is running a number of libvirt / kvm virtuals on rbd via librados: Around the time the "slow requests" messages pop up, the virtual machines experience a spike in disk latency, e.g. from inside one of the virtuals: load %user %nice %sys %iow %stl %idle dev rrqm/s wrqm/s r/sw/srkB/swkB/s arq-sz aqu-sz await rwait wwait %util 2013-02-12-18:55:23 1.7 0.3 0.0 0.9 42.2 0.0 18.5 vdb0.0 25.8 0.09.6 0.00 136.53 28.44 1.34 139.56 0.00 139.6 70.7 2013-02-12-18:55:38 1.5 0.9 0.0 0.4 35.8 0.0 42.8 vdb0.0 26.6 0.0 10.8 0.00 143.20 26.52 0.46 41.98 0.00 42.0 40.0 2013-02-12-18:55:53 1.4 0.7 0.0 0.4 38.0 0.0 51.7 vdb0.07.5 0.06.8 0.0053.33 15.69 0.46 68.08 0.00 68.1 38.8 2013-02-12-18:56:10 2.2 0.0 0.0 0.1 8.0 0.0 15.8 vdb0.02.1 0.00.2 0.00 9.18 78.00 0.98 4164.00 0.00 4164.0 70.6 2013-02-12-18:56:32 3.7 0.0 0.0 0.1 0.0 0.0 0.0 vdb0.00.1 0.00.8 0.00 3.27 8.00 24.12 14519.78 0.00 14519.8 100.0 2013-02-12-18:56:47 5.1 0.5 0.0 0.6 0.0 0.0 0.0 vdb0.0 11.5 0.05.4 0.0065.87 24.40 1.60 3620.15 0.00 3620.1 88.4 2013-02-12-18:57:03 5.2 0.6 0.0 1.1 3.5 0.0 0.0 vdb0.0 19.6 0.06.9 0.00 101.25 29.19 1.13 162.41 0.00 162.4 87.4 2013-02-12-18:57:20 5.2 0.6 0.0 0.9 9.7 0.0 0.0 vdb0.0 44.6 0.09.7 0.00 214.82 44.27 2.41 248.22 0.00 248.2 85.3 2013-02-12-18:57:36 4.4 0.4 0.0 0.5 28.7 0.0 46.5 vdb0.0 17.6 0.05.2 0.0087.50 33.33 0.56 107.14 0.00 107.1 48.5 ...corresponding to this in the b5 / osd.1 log: 2013-02-12 18:52:08.812880 7f9c0bc16700 0 monclient: hunting for new mon 2013-02-12 18:55:18.851791 7f9c0bc16700 0 monclient: hunting for new mon 2013-02-12 18:56:25.414948 7f9c13c26700 0 log [WRN] : 6 slow requests, 6 included below; oldest blocked for > 30.372124 secs 2013-02-12 18:56:25.414958 7f9c13c26700 0 log [WRN] : slow request 30.372124 seconds old, received at 2013-02-12 18:55:55.042767: osd_op(client.9821.0:144779 rb.0.209f.74b0dc51.0023 [write 1593344~4096] 2.1882ddb7) v4 currently no flag points reached ... 2013-02-12 18:57:13.427008 7f9c13c26700 0 log [WRN] : slow request 40.721769 seconds old, received at 2013-02-12 18:56:32.705190: osd_op(client.9821.0:146756 rb.0.209f.74b0dc51.0128 [write 819200~8192] 2.b4390173) v4 currently commit sent 2013-02-12 18:59:43.886517 7f9c0bc16700 0 monclient: hunting for new mon 2013-02-12 19:02:53.911641 7f9c0bc16700 0 monclient: hunting for new mon Cheers, Chris. -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html