As nautilus is nearing its EOL we are planning to do a next (maybe the
last) point release - nautilus 14.2.22 in the first part of June 2021.
If you have any code changes that are to be included pls raise PRs and
add labels "nautilus -batch1" and "needs-qa", so they can be tested
and merged in time for 14.2.22.
With Pacific (16.2.0) out of the door we have the RBD persistent
WriteBack cache for RBD:
Has anybody performed some benchmarks with the RBD cache?
Interested in the QD=1 bs=4k performance mainly.
I don't have any proper hardware available to run benchmarks on yet.
We have two issues:
Caused by numpy not supporting Python sub-interpreters. Unfortunately, the
latter issue came up in the most recent Octopus validations. I suspect
a matter of time, till our users are affected by it.
Note that removing numpy is not easy, as kuberenetes-client depends
(transitively) on numpy.
The performance meeting will be starting in about 70 minutes. Today the
only topic I have is that we'll discuss whether or not it makes sense to
do another submission for CephFS for the ISC21 IO500 competition.
Please feel free to add your own topic!
Highlights from this week:
* Ceph User Survey '21: Dashboard analysis
* Some responses looked contradictory (responding not using the
dashboard and then reporting a 5/5 rating)
* Would you recommend dashboard?
- average 6.75
- median 8
* Does the dashboard allow you to perform tasks faster than cli?
- No 52
- Yes 48
* What's most used in the dashboard
- landing page
- iscsi and NFS least used
- RBD > Cephfs consistent Ceph-wise usage
- RGW should rank higher than cephfs 62% vs 54%
* Detractors still use the landing page, cluster logs, mon status, and
* Dashboard user Profiles
- Haven't used the dashboard yet
- Older versions of Ceph Dashboard
- Documentation issues?
- Getting started doc refers to the CLI
- Performance concerns?
* monitoring only
- Read-only mode
- Fear of breaking things?
* CLI junkies
- Claim: "CLI is faster, easy to remember and scriptable
- Dashboard provides bulk ops and multi-step workflows
- API recorder in OpenAttic:
- Dashboard maps 1:1 to Ceph CLI
* other mgmt tools
- Using Ceph-ansible, Puppet, Rook
- Focus efforts on the landing page and cluster logs
- Increase inline docs and contextual helpers
- Add support for bulk ops
- Rest API recorder: generate a script from UI actions
- Publicize dashboard in Ceph blog, tech talks.
* Ceph.io website updates
- The working group's ideal launch date is June 24th
- Content needed for pages
* Market Development Working Group
- The group would like to be more connected with the CLT on the
roadmap for releases.
- they meet once a month.
- they're interested in roadmaps, blog posts,
- Mike: send out pad for gathering high level features/improvements
in Quincy roadmap
* NFS questions on dev@ need answers...
- ganesha and dir_chunk=0 - problematic for RGW exports
* GSOC onboarding / collaboration
- ali to organize
octopus: subinterpreters again
- Planned to be discussed at the next CDM