Re: Request for Info: bluestore_compression_mode?

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

 



Hi, maybe you remember this line in my long report:

> bluestore_compression_min_blob_size_ssd = 8192 *** Dang! ***

To fix the issues with settings on our cluster, I plan to change the bluestore_compression_min_blob_size_ssd and also disable compression on pools that are accidentally compressed due to bluestore settings, but shouldn't. This is the plan:

ceph config set osd bluestore_compression_min_blob_size_ssd = 65536

ceph osd pool set sr-rbd-meta-one compression_mode none
ceph osd pool set con-fs2-meta1 compression_mode none
ceph osd pool set con-fs2-meta2 compression_mode none
ceph osd pool set sr-rbd-data-one-perf compression_mode none

To be absolutely sure, could you please confirm that these changes can be executed at run time on a system that has data on it?

Thanks a lot!
=================
Frank Schilder
AIT Risø Campus
Bygning 109, rum S14

________________________________________
From: Frank Schilder
Sent: 31 August 2022 11:32
To: Mark Nelson; ceph-users@xxxxxxx
Cc: Adam Kupczyk
Subject: Re:  Request for Info: bluestore_compression_mode?

Hi Mark,

I need a bit more time for replying. I just wanted to add a few observations and correct some mistakes in my earlier communication.

- There are actually 5 different settings, NULL (=unset), none, passive, aggressive and force. These can be set on OSDs and pools. In addition, the blob-size arguments for compression are also available on pools. This gets really confusing and it is not fully documented which settings will take precedence.

- In Octopus it seems that the interaction of combinations of OSD- and pool settings has been fixed (at least the case where one of these is NULL). This was a mess in mimic. Its still not clear from the documentation what takes precedence in all possible situations of conflicts.

My personal take is that pool-setings are sufficient, the OSD settings could be removed and the mode options could be reduced to none and force, where NULL=none (none is default).

As promised, I will return to this in another thread about the small files/tail object handling that will also consider compression.

Best regards,
=================
Frank Schilder
AIT Risø Campus
Bygning 109, rum S14
_______________________________________________
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