- ceph-users - lists.ceph.io

Re: Re layout help: need chassis local io to minimize net links
by Anthony D'Atri 30 Jun '20

30 Jun '20

What does “traffic” mean? Reads? Writes will have to hit the net regardless of any machinations. > On Jun 29, 2020, at 7:31 PM, Harry G. Coin <hgcoin(a)gmail.com> wrote: > > I need exactly what ceph is for a whole lot of work, that work just > doesn't represent a large fraction of the total local traffic. Ceph is > the right choice. Plainly ceph has tremendous support for replication > within a chassis, among chassis and among racks. I just need > intra-chassis traffic to not hit the net much. Seems not such an > unreasonable thing given the intra-chassis crush rules and all. After > all.. ceph's name wasn't chosen for where it can't go.... > >>>> On 6/29/20 1:57 PM, Marc Roos wrote: >> I wonder if you should not have chosen a different product? Ceph is >> meant to distribute data across nodes, racks, data centers etc. For a >> nail use a hammer, for a screw use a screw driver. >> -----Original Message----- >> To: ceph-users(a)ceph.io >> Subject: *****SPAM***** [ceph-users] layout help: need chassis local io >> to minimize net links >> Hi >> I have a few servers each with 6 or more disks, with a storage workload >> that's around 80% done entirely within each server. From a >> work-to-be-done perspective there's no need for 80% of the load to >> traverse network interfaces, the rest needs what ceph is all about. So >> I cooked up a set of crush maps and pools, one map/pool for each server >> and one map/pool for the whole. Skipping the long story, the >> performance remains network link speed bound and has got to change. >> "Chassis local" io is too slow. I even tried putting a mon within each >> server. I'd like to avoid having to revert to some other HA >> filesystem per server with ceph at the chassis layer if I can help >> it. >> Any notions that would allow 'chassis local' rbd traffic to avoid or >> mostly avoid leaving the box? >> Thanks! >> _______________________________________________ >> ceph-users mailing list -- ceph-users(a)ceph.io To unsubscribe send an >> email to ceph-users-leave(a)ceph.io > _______________________________________________ > ceph-users mailing list -- ceph-users(a)ceph.io > To unsubscribe send an email to ceph-users-leave(a)ceph.io

2 3

Re: Re layout help: need chassis local io to minimize net links
by Jeff Welling 30 Jun '20

30 Jun '20

I've not been replying to the list, apologies. > just the write metadata to the mon, with the actual write data content not having to cross a physical ethernet cable but directly to the chassis-local osds via the 'virtual' internal switch? This is my understanding as well, yes. I've not explored the ceph source yet though. On 2020-06-29 8:37 p.m., Harry G. Coin wrote: > > Jeff, thanks for the lead. When a user space rbd write has as a > destination three replica osds in the same chassis, does the whole > write get shipped out to the mon and then back, or just the write > metadata to the mon, with the actual write data content not having to > cross a physical ethernet cable but directly to the chassis-local osds > via the 'virtual' internal switch? I thought when I read the layout > of how ceph works only the control traffic goes to the mons, the data > directly from the generator to the osds. Did I get that wrong? > > > On 6/29/20 10:32 PM, Jeff W wrote: >> You mentioned setting up pools per host but still hitting network >> limits, did you try tcpdumping the NIC to see who's talking to who? >> Perhaps something isn't configured the way you expect? That may help >> you narrow down what is using the NIC as well, Mon or osd or what >> not. If it's local, I would think that the NIC wouldn't be a >> bottleneck and if it is a bottleneck I would suspect my own configs, >> but that's just my 2c. >> >> Off the top of my head im thinking it's the Mon, because even if you >> setup multiple pools I can't think of a way to have multiple groups >> of mons maintaining their own shards of consensus. Unless your >> workload is largely read only, then .. I'm not sure what the >> bottleneck would be. >> >> >> On Mon., Jun. 29, 2020, 7:32 p.m. Harry G. Coin, <hgcoin(a)gmail.com >> <mailto:hgcoin@gmail.com>> wrote: >> >> I need exactly what ceph is for a whole lot of work, that work just >> doesn't represent a large fraction of the total local traffic. >> Ceph is >> the right choice. Plainly ceph has tremendous support for >> replication >> within a chassis, among chassis and among racks. I just need >> intra-chassis traffic to not hit the net much. Seems not such an >> unreasonable thing given the intra-chassis crush rules and all. >> After >> all.. ceph's name wasn't chosen for where it can't go.... >> >> On 6/29/20 1:57 PM, Marc Roos wrote: >> > I wonder if you should not have chosen a different product? >> Ceph is >> > meant to distribute data across nodes, racks, data centers etc. >> For a >> > nail use a hammer, for a screw use a screw driver. >> > >> > >> > -----Original Message----- >> > To: ceph-users(a)ceph.io <mailto:ceph-users@ceph.io> >> > Subject: *****SPAM***** [ceph-users] layout help: need chassis >> local io >> > to minimize net links >> > >> > Hi >> > >> > I have a few servers each with 6 or more disks, with a storage >> workload >> > that's around 80% done entirely within each server. From a >> > work-to-be-done perspective there's no need for 80% of the load to >> > traverse network interfaces, the rest needs what ceph is all >> about. So >> > I cooked up a set of crush maps and pools, one map/pool for >> each server >> > and one map/pool for the whole. Skipping the long story, the >> > performance remains network link speed bound and has got to >> change. >> > "Chassis local" io is too slow. I even tried putting a mon >> within each >> > server. I'd like to avoid having to revert to some other HA >> > filesystem per server with ceph at the chassis layer if I can help >> > it. >> > >> > Any notions that would allow 'chassis local' rbd traffic to >> avoid or >> > mostly avoid leaving the box? >> > >> > Thanks! >> > >> > >> > >> > >> > _______________________________________________ >> > ceph-users mailing list -- ceph-users(a)ceph.io >> <mailto:ceph-users@ceph.io> To unsubscribe send an >> > email to ceph-users-leave(a)ceph.io <mailto:ceph-users-leave@ceph.io> >> > >> > >> _______________________________________________ >> ceph-users mailing list -- ceph-users(a)ceph.io >> <mailto:ceph-users@ceph.io> >> To unsubscribe send an email to ceph-users-leave(a)ceph.io >> <mailto:ceph-users-leave@ceph.io> >>

1 0

Debian install
by Rafael Quaglio 30 Jun '20

30 Jun '20

Hi, We have already installed a new Debian (10.4) server and I need put it in a Ceph cluster. When I execute the command to install ceph on this node: ceph-deploy install --release nautilus node1 It starts to install a version 12.x in my node... (...) [serifos][DEBUG ] After this operation, 183 MB of additional disk space will be used. [serifos][DEBUG ] Selecting previously unselected package python-cephfs. (Reading database ... 30440 files and directories currently installed.) [serifos][DEBUG ] Preparing to unpack .../python-cephfs_12.2.11+dfsg1-2.1+b1_amd64.deb ... [serifos][DEBUG ] Unpacking python-cephfs (12.2.11+dfsg1-2.1+b1) ... [serifos][DEBUG ] Selecting previously unselected package ceph-common. [serifos][DEBUG ] Preparing to unpack .../ceph-common_12.2.11+dfsg1-2.1+b1_amd64.deb ... [serifos][DEBUG ] Unpacking ceph-common (12.2.11+dfsg1-2.1+b1) ... (...) How do I upgrade this packages? Even installed packages in this version, the installation completes without erros. The question is due to an error message that I'm recieving when deploy a new osd. ceph-deploy osd create --data /dev/sdb node1 At this point: [ceph_deploy.osd][INFO ] Distro info: debian 10.4 buster [ceph_deploy.osd][DEBUG ] Deploying osd to node1 [node1][DEBUG ] write cluster configuration to /etc/ceph/{cluster}.conf [node1][DEBUG ] find the location of an executable [node1][INFO ] Running command: sudo /usr/sbin/ceph-volume --cluster ceph lvm create --bluestore --data /dev/sdb [node1][WARNIN] --> RuntimeError: Unable to create a new OSD id [node1][DEBUG ] Running command: /bin/ceph-authtool --gen-print-key [node1][DEBUG ] Running command: /bin/ceph --cluster ceph --name client.bootstrap-osd --keyring /var/lib/ceph/bootstrap-osd/ceph.keyring -i - osd new 76da6c51-8385-4ffc-9a8e-0dfc11e31feb [node1][DEBUG ] stderr: /build/ceph-qtARip/ceph-12.2.11+dfsg1/src/mon/MonMap.cc: In function 'void MonMap::sanitize_mons(std::map<std::__cxx11::basic_string<char>, entity_addr_t>&)' thread 7f2bc7fff700 time 2020-06-29 06:56:17.331350 [node1][DEBUG ] stderr: /build/ceph-qtARip/ceph-12.2.11+dfsg1/src/mon/MonMap.cc: 77: FAILED assert(mon_info[p.first].public_addr == p.second) [node1][DEBUG ] stderr: ceph version 12.2.11 (26dc3775efc7bb286a1d6d66faee0ba30ea23eee) luminous (stable) [node1][DEBUG ] stderr: 1: (ceph::__ceph_assert_fail(char const*, char const*, int, char const*)+0xf5) [0x7f2bdaff5f75] [node1][DEBUG ] stderr: 2: (MonMap::sanitize_mons(std::map<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >, entity_addr_t, std::less<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const, entity_addr_t> > >&)+0x568) [0x7f2bdb050038] [node1][DEBUG ] stderr: 3: (MonMap::decode(ceph::buffer::list::iterator&)+0x4da) [0x7f2bdb05500a] [node1][DEBUG ] stderr: 4: (MonClient::handle_monmap(MMonMap*)+0x216) [0x7f2bdb042a06] [node1][DEBUG ] stderr: 5: (MonClient::ms_dispatch(Message*)+0x4ab) [0x7f2bdb04729b] [node1][DEBUG ] stderr: 6: (DispatchQueue::entry()+0xeba) [0x7f2bdb06bf5a] [node1][DEBUG ] stderr: 7: (DispatchQueue::DispatchThread::entry()+0xd) [0x7f2bdb1576fd] [node1][DEBUG ] stderr: 8: (()+0x7fa3) [0x7f2be499dfa3] [node1][DEBUG ] stderr: 9: (clone()+0x3f) [0x7f2be45234cf] [node1][DEBUG ] stderr: NOTE: a copy of the executable, or `objdump -rdS <executable>` is needed to interpret this. [node1][ERROR ] RuntimeError: command returned non-zero exit status: 1 [ceph_deploy.osd][ERROR ] Failed to execute command: /usr/sbin/ceph-volume --cluster ceph lvm create --bluestore --data /dev/sdb [ceph_deploy][ERROR ] GenericError: Failed to create 1 OSDs I think this error occurs because the wrong package that was installed. Thanks, Rafael

4 4

Octopus missing rgw-orphan-list tool
by Andrei Mikhailovsky 29 Jun '20

29 Jun '20

Hello, I have been struggling a lot with radosgw buckets space wastage, which is currently stands at about 2/3 of utilised space is wasted and unaccounted for. I've tried to use the tools to find the orphan objects, but these were running in loop for weeks on without producing any results. Wido and a few others pointed out that this function is broke in was deprecated and that instead the rgw-orphan-list should be used instead. I have upgraded to Octopus and I have been following the documentation [ https://docs.ceph.com/docs/master/radosgw/orphans/ | https://docs.ceph.com/docs/master/radosgw/orphans/ ] . However, the ceph and radon packages for Ubuntu 18.04 do not seem to have this tool. The same applies to the bucket radoslist option to the radosgw-admin command. root@arh-ibstorage1-ib:~# radosgw-admin bucket radoslist ERROR: Unrecognized argument: 'radoslist' Expected one of the following: check chown limit link list reshard rewrite rm stats sync unlink root@arh-ibstorage1-ib:~# dpkg -l *rados\* Desired=Unknown/Install/Remove/Purge/Hold | Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend |/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad) ||/ Name Version Architecture Description +++-=======================-================-================-==================================================== un librados <none> <none> (no description available) ii librados2 15.2.3-1bionic amd64 RADOS distributed object store client library ii libradosstriper1 15.2.3-1bionic amd64 RADOS striping interface ii python3-rados 15.2.3-1bionic amd64 Python 3 libraries for the Ceph librados library ii radosgw 15.2.3-1bionic amd64 REST gateway for RADOS distributed object store I am running Ubuntu 18.04 with version 15.2.3 of ceph and radosgw. Please suggest what should I do to remove the wasted space that radosgw is creating? I've calculated the wasted space by adding up the reported usage of all the buckets and checking it agains the output of the rados df command. The buckets are using around 11TB. The rados df reports 68TB of usage with replica of 2. Rather alarming! Thanks for you help

1 0

Nautilus latest builds for CentOS 8
by Victoria Martinez de la Cruz 29 Jun '20

29 Jun '20

Hi all, I'm wondering if there is a plan to add Nautilus builds for CentOS 8 [0]. Right now I see there are builds for CentOS 7, but for CentOS 8 there are only builds for Octopus and master. Thanks, V [0] https://shaman.ceph.com/api/search/?project=ceph&distros=centos/8&ref=nauti…

6 14

Octopus Grafana inside the dashboard
by Simon Sutter 29 Jun '20

29 Jun '20

Hello I'm trying to get Grafana working inside the Dashboard. If I press on "Overall Performance" tab, I get an error, because the iframe tries to connect to the internal hostname, which cannot be resolved from my machine. If I directly open grafana, everything works. How can I tell the dashboard, to use the full domain name? I have tried to set "ceph dashboard set-grafana-api-url https://node01.mycorp.local:3000" but that does not work and always sets itself back to "https://node01:3000". Thanks in advance, Simon

2 1

Nautilus 14.2.10 mon_warn_on_pool_no_redundancy
by Wout van Heeswijk 29 Jun '20

29 Jun '20

Hi All, I really like the idea of warning users against using unsafe practices. Wouldn't it make sense to warn against using min_size=1 instead of size=1. I've seen data loss happen with size=2 min_size=1 when multiple failures occur and write have been done between the failures. Effectively the new warning below says "It is not considered safe to run with no redundancy". Which is true, but when failure occurs or maintenance is executed, with size=2 and min_size=1, as soon as data is written, there might not be data redundancy for that newly written data. A failure of an OSD at that moment would result in data loss. Since you cannot run size=1 with min_size > 1, this use-case would also be covered. I understand this has implications for size=2 when executing maintenance, but I think most people are not aware of the risks they are taking with min_size=1. Those that are aware can suppress the warning. * Ceph will issue a health warning if a RADOS pool's `size` is set to 1 or in other words the pool is configured with no redundancy. This can be fixed by setting the pool size to the minimum recommended value with:: ceph osd pool set <pool-name> size <num-replicas> The warning can be silenced with:: ceph config set global mon_warn_on_pool_no_redundancy false -- kind regards, Wout 42on

2 1

Re: rgw : unable to find part(s) of aborted multipart upload of [object].meta
by EDH - Manuel Rios 29 Jun '20

29 Jun '20

Hi Amit, Yes in the non-ec pool there’re like 600 files .meta but i dont know if its safe move to data pool. Does anyone know if there is a way to generate a synthetic .meta only so that the delete command is capable of deleting the file? Regards Manuel De: Amit Ghadge <amitg.b14(a)gmail.com> Enviado el: lunes, 29 de junio de 2020 6:14 Para: EDH - Manuel Rios <mriosfer(a)easydatahost.com> Asunto: Re: [ceph-users] rgw : unable to find part(s) of aborted multipart upload of [object].meta You can also check in default.rgw.buckets.non-ec pool for unmerged multipart or bucket indexes that need to be fix. On Mon, Jun 29, 2020 at 5:56 AM EDH - Manuel Rios <mriosfer(a)easydatahost.com<mailto:mriosfer@easydatahost.com>> wrote: Hi Dev's With the failures of the previous versions in the buckets due to the shardings. We have started a copy of the buckets to new buckets to clean our ceph cluster. After synchronizing the bucket with the AWS cli, we are in the phase of deleting the old buckets. We have tried unsuccessfully: radosgw-admin bucket rm --bucket = XXXX --purge-objects Result is a loop that shows with the debug: "NOTE: unable to find part (s) of aborted multipart upload of [object] .meta" After seeing this failure we have tried to clean up using "rados". For this we have listed all the objects belonging to the pool bucket using its marker_id. Once done, we send them a script to delete massively, with "rados -p [rgw-pool.data] rm [object]" The result of all is similar to the following: rados -p default.rgw.buckets.data rm 48efb8c3-693c-4fe0-bbe4-fdc16f590a82.3886182.18__multipart_MBS-3369403d-e0bf-45e3-89ba-614b6d390dc5/CBB_BIM-EURODG/CBB_DiskImage/Disk_00000000-0000-0000-0000-000000000000/Volume_NTFS_00000000-0000-0000-0000-000000000001$/20200104230152/131.cbrevision.5K5_leiUZoHQsjBvUxw2QbM1WQPQLlc error removing default.rgw.buckets.data>48efb8c3-693c-4fe0-bbe4-fdc16f590a82.3886182.18__multipart_MBS-3369403d-e0bf-45e3-89ba-614b6d390dc5/CBB_BIM-EURODG/CBB_DiskImage/Disk_00000000-0000-0000-0000-000000000000/Volume_NTFS_00000000-0000-0000-0000-000000000001$/20200104230152/131.cbrevision.5K5_leiUZoHQsjBvUxw2QbM1WQPQLlc: (2) No such file or directory Any idea or better way to clean the cluster of these objects? Our estimates are 100TB of wrong objects. Regards Manuel _______________________________________________ ceph-users mailing list -- ceph-users(a)ceph.io<mailto:ceph-users@ceph.io> To unsubscribe send an email to ceph-users-leave(a)ceph.io<mailto:ceph-users-leave@ceph.io>

1 0

mgr log shows a lot of ms_handle_reset messages
by XuYun 29 Jun '20

29 Jun '20

Hi, We are running Ceph nautilus (14.2.10 now), every mgr reports a ‘ms_handle_reset’ message every second. Would it be a normal behavior? 2020-06-28 16:35:43.580 7fcdc96e6700 0 client.0 ms_handle_reset on v2:111.111.121.3:6807/7 2020-06-28 16:35:44.580 7fcdc96e6700 0 client.0 ms_handle_reset on v2:111.111.121.3:6807/7 2020-06-28 16:35:45.580 7fcdc96e6700 0 client.0 ms_handle_reset on v2:111.111.121.3:6807/7 2020-06-28 16:35:46.581 7fcdc96e6700 0 client.0 ms_handle_reset on v2:111.111.121.3:6807/7 2020-06-28 16:35:47.580 7fcdc96e6700 0 client.0 ms_handle_reset on v2:111.111.121.3:6807/7 Br, Xu Yun

2 4

rgw : unable to find part(s) of aborted multipart upload of [object].meta
by EDH - Manuel Rios 29 Jun '20

29 Jun '20

Hi Dev's With the failures of the previous versions in the buckets due to the shardings. We have started a copy of the buckets to new buckets to clean our ceph cluster. After synchronizing the bucket with the AWS cli, we are in the phase of deleting the old buckets. We have tried unsuccessfully: radosgw-admin bucket rm --bucket = XXXX --purge-objects Result is a loop that shows with the debug: "NOTE: unable to find part (s) of aborted multipart upload of [object] .meta" After seeing this failure we have tried to clean up using "rados". For this we have listed all the objects belonging to the pool bucket using its marker_id. Once done, we send them a script to delete massively, with "rados -p [rgw-pool.data] rm [object]" The result of all is similar to the following: rados -p default.rgw.buckets.data rm 48efb8c3-693c-4fe0-bbe4-fdc16f590a82.3886182.18__multipart_MBS-3369403d-e0bf-45e3-89ba-614b6d390dc5/CBB_BIM-EURODG/CBB_DiskImage/Disk_00000000-0000-0000-0000-000000000000/Volume_NTFS_00000000-0000-0000-0000-000000000001$/20200104230152/131.cbrevision.5K5_leiUZoHQsjBvUxw2QbM1WQPQLlc error removing default.rgw.buckets.data>48efb8c3-693c-4fe0-bbe4-fdc16f590a82.3886182.18__multipart_MBS-3369403d-e0bf-45e3-89ba-614b6d390dc5/CBB_BIM-EURODG/CBB_DiskImage/Disk_00000000-0000-0000-0000-000000000000/Volume_NTFS_00000000-0000-0000-0000-000000000001$/20200104230152/131.cbrevision.5K5_leiUZoHQsjBvUxw2QbM1WQPQLlc: (2) No such file or directory Any idea or better way to clean the cluster of these objects? Our estimates are 100TB of wrong objects. Regards Manuel

1 0