Re: Getting started with cephfs-top, how to install

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Thank you for the reply, Neeraj - solved!

I was just going off of the document at https://docs.ceph.com/en/quincy/cephfs/cephfs-top/ - I did not see a specified terminal emulator listed, but yes, after switching my Secure Shell program to XTerm, that seems to have fixed it. Also, now 'ceph fs perf stats' shows a "wall of text" in an unformatted way, which is what I assume cephfs-top is formatting so nicely in its python script.

I appreciate the help, and I will make an edit request to that document adding the information I have learned from you and Xiubo, so other relative novices like myself will perhaps be able to fix these issues on their own in the future!

Best,

Zach Heise


On 2022-10-19 2:16 PM, Neeraj Pratap Singh wrote:
Hi Zach,
Seeing the `fs perf stats` output , it looks like you are not using the latest build. Lots of enhancements are being done in cephfs-top recently.
Will suggest to use latest build for better results.
And regarding your `use_default_colors()` error, it looks like there some issue with the unsupported terminal emulator.

On Thu, Oct 20, 2022 at 12:20 AM Zach Heise (SSCC) <heise@xxxxxxxxxxxx> wrote:

Thank you, Xiubo - yes, checking my ceph.repo file as specified at https://docs.ceph.com/en/pacific/install/get-packages/#rhel reminded me that I had set the ceph-noarch repo to disabled, because we didn't want ceph trying to update itself outside of using cephadm for new builds. Flipping that bit to 1 then re-running yum update then yum install cephfs-top solved that problem.

However, that leads to a second question then - after running the required command to create the fstop user with $ ceph auth get-or-create client.fstop mon 'allow r' mds 'allow r' osd 'allow r' mgr 'allow r' as the instructions specify, and adding the file ceph.client.fstop.keyring with the relevant key to /etc/ceph on the node in question (which the instructions do not specify and I will make a github edit later), I am getting an error:

"exception: use_default_colors() returned ERR"

Beyond that, it seems like the system that cephfs-top needs to work, perf stats, has no data in the output fields?

ceph01.ssc.wisc.edu> ceph fs perf stats
{"version": 1, "global_counters": ["cap_hit", "read_latency", "write_latency", "metadata_latency", "dentry_lease", "opened_files", "pinned_icaps", "opened_inodes", "read_io_sizes", "write_io_sizes"], "counters": [], "client_metadata": {}, "global_metrics": {}, "metrics": {"delayed_ranks": []}}

I activated ceph fs perf stats yesterday, so by this point I should have data in the stats, unless there is a problem elsewhere?

Zach Heise


On 2022-10-18 7:56 PM, Xiubo Li wrote:

Hi Zach,

On 18/10/2022 04:20, Zach Heise (SSCC) wrote:

I'd like to see what CephFS clients are doing the most IO. According to this page: https://docs.ceph.com/en/quincy/cephfs/cephfs-top/ - cephfs-top is the simplest way to do this? I enabled 'ceph mgr module enable stats' today, but I'm a bit confused about what the best way is to get the cephfs-top package to use this perf stats.

If you are building code from source then you should run it just by:

[ceph/build]$ ./src/tools/cephfs/top/cephfs-top

You can also find the main/quincy/pacific builds for different distro pacakges under [1] and their artifacts are made available under [2].

[1] https://shaman.ceph.com/builds/ceph/
[2] https://shaman.ceph.com/repos/ceph/

Thanks!

- Xiubo


The ceph doc page linked above just mentions "cephfs-top is available as part of cephfs-top package" but it does not list what repo is required to access this. Anyone using cephfs-top themselves, and know the missing parts of this document that should be added?

--
Zach Heise

_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx


--
Neeraj Pratap Singh
He/Him/His
Associate Software Engineer, CephFS
neesingh@xxxxxxxxxx

Red Hat Inc.



_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx

[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Ceph Dev]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux