How to test run the new EC Erasure Coding Enhancements Optimizations?
Hi, I am new ignorant user and want to practice test any of the new EC Optimizations that are being worked on for tentacle. How can I test install and setup using any of the new code that includes the EC Optimizations? What would be easiest steps to standup a test environment and how to run pre-release code? Thanks
Hi, The new code is in the main branch and tentacle branch, there will be a tentacle release candidate very soon, but not quite yet so you’ll probably have to build it from source now or wait for the release candidate. To use the new feature, you need to enable a new flag on your EC pool: ceph osd pool set {pool-name} allow_ec_optimizations on If you are creating a new pool, the recommendation is to increase the stripe unit from the current default of 4K to 16K – you can either set this in the erasure code profile or by changing the conf file setting osd_pool_erasure_code_stripe_unit. See this PR https://github.com/ceph/ceph/pull/64809 which has the documentation updates for the new feature for more information. Cheers, Bill. bill_scales@uk.ibm.com<mailto:bill_scales@uk.ibm.com> IBM Distinguished Engineer, IBM Storage From: Chris Lawsonn <chrislawsonn@outlook.com> Date: Monday, 4 August 2025 at 14:43 To: ceph-users@ceph.io <ceph-users@ceph.io> Subject: [EXTERNAL] [ceph-users] How to test run the new EC Erasure Coding Enhancements Optimizations? Hi, I am new ignorant user and want to practice test any of the new EC Optimizations that are being worked on for tentacle. How can I test install and setup using any of the new code that includes the EC Optimizations? What would be easiest steps to standup a test environment and how to run pre-release code? Thanks _______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io Unless otherwise stated above: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: Building C, IBM Hursley Office, Hursley Park Road, Winchester, Hampshire SO21 2JN
I haven't tried the upmap-read balancer myself, but have you removed the upmap entries before trying to revert? ceph osd rm-pg-upmap-primary-all https://docs.ceph.com/en/reef/rados/operations/read-balancer/#removing-pg-up... Zitat von quaglio@bol.com.br:
I enabled the upmap-read balancemant, but in my environment, it is no longer possible to mount a ceph filesystem. I tried to revert to the old settings, but I get the following message in logs:
[Mon Aug 11 11:42:05 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:05 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon0 (1)192.168.193.151:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon0 (1)192.168.193.151:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:07 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features [Mon Aug 11 11:42:09 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:09 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features
I can't mount any more ceph filesystems. The ceph processes are up.
# uname -a Linux k02 5.15.0-141-generic #151-Ubuntu SMP Sun May 18 21:35:19 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
I've already tried to revert to the old settings, but when I try to do so, I get the following alert: # ceph osd set-require-min-compat-client luminous Error EPERM: osdmap current utilizes features that require reef; cannot set require_min_compat_client below that to luminous
Could you help? Enviado através do Bol Mail[1]
Links: ------ [1] https://produtos.uol.com.br/bol
Hello, If you're seeing this error, it means that there are clients in your cluster that do not support Reef features, which `pg-upmap-primary` is. To get rid of the errors, you have to change the balancer back to upmap mode, then remove the pg-upmap-primary mappings: ceph balancer mode upmap ceph osd rm-pg-upmap-primary-all This behavior is expected when older clients try to connect to the cluster with newer (in this case, Reef) features. I suspect that you enabled upmap-read mode before mounting the filesystem, since enabling upmap-read mode requires you to set "require-min-compat-client=reef", which is basically a gate to prevent older clients from connecting to a cluster with Reef features. Can you confirm that this was the order of operations? Also, please confirm that running the above commands gets rid of the errors. Thanks, Laura On Mon, Aug 11, 2025 at 2:38 PM Eugen Block <eblock@nde.ag> wrote:
I haven't tried the upmap-read balancer myself, but have you removed the upmap entries before trying to revert?
ceph osd rm-pg-upmap-primary-all
https://docs.ceph.com/en/reef/rados/operations/read-balancer/#removing-pg-up...
Zitat von quaglio@bol.com.br:
I enabled the upmap-read balancemant, but in my environment, it is no longer possible to mount a ceph filesystem. I tried to revert to the old settings, but I get the following message in logs:
[Mon Aug 11 11:42:05 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:05 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon0 (1)192.168.193.151:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon0 (1)192.168.193.151:6789 missing required protocol features [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:06 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] ceph: No mds server is up or the cluster is laggy [Mon Aug 11 11:42:07 2025] libceph: mon2 (1)192.168.193.153:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:07 2025] libceph: mon2 (1)192.168.193.153:6789 missing required protocol features [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:08 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features [Mon Aug 11 11:42:09 2025] libceph: mon1 (1)192.168.193.152:6789 feature set mismatch, my 2f018fb87aa4aafe < server's 2f018fb8faa4aafe, missing 80000000 [Mon Aug 11 11:42:09 2025] libceph: mon1 (1)192.168.193.152:6789 missing required protocol features
I can't mount any more ceph filesystems. The ceph processes are up.
# uname -a Linux k02 5.15.0-141-generic #151-Ubuntu SMP Sun May 18 21:35:19 UTC 2025 x86_64 x86_64 x86_64 GNU/Linux
I've already tried to revert to the old settings, but when I try to do so, I get the following alert: # ceph osd set-require-min-compat-client luminous Error EPERM: osdmap current utilizes features that require reef; cannot set require_min_compat_client below that to luminous
Could you help? Enviado através do Bol Mail[1]
Links: ------ [1] https://produtos.uol.com.br/bol
_______________________________________________ ceph-users mailing list -- ceph-users@ceph.io To unsubscribe send an email to ceph-users-leave@ceph.io
-- Laura Flores She/Her/Hers Software Engineer, Ceph Storage <https://ceph.io> Chicago, IL lflores@ibm.com | lflores@redhat.com <lflores@redhat.com> M: +17087388804
participants (5)
-
Bill Scales
-
Chris Lawsonn
-
Eugen Block
-
Laura Flores
-
quaglio＠bol.com.br