13 May
2020
13 May
'20
2:09 p.m.
Hi
On one of our Ceph clusters, some OSDs have been marked as full. Since this is a staging
cluster that does not have much data on it, this is strange.
Looking at the full OSDs through “ceph osd df” I figured out that the space is mostly used
by metadata:
SIZE: 122 GiB
USE: 118 GiB
DATA: 2.4 GiB
META: 116 GiB
We run mimic, and for the affected OSDs we use a db device (nvme) in addition to the
primary device (hdd).
In the logs we see the following errors:
2020-05-12 17:10:26.089 7f183f604700 1 bluefs _allocate failed to allocate 0x400000
on bdev 1, free 0x0; fallback to bdev 2
2020-05-12 17:10:27.113 7f183f604700 1 bluestore(/var/lib/ceph/osd/ceph-8)
_balance_bluefs_freespace gifting 0x180a000000~400000 to bluefs
2020-05-12 17:10:27.153 7f183f604700 1 bluefs add_block_extent bdev 2
0x180a000000~400000
We assume it is an issue with Rocksdb, as the following call will quickly fix the
problem:
ceph daemon osd.8 compact
The question is, why is this happening? I would think that “compact" is something
that runs automatically from time to time, but I’m not sure.
Is it on us to run this regularly?
Any pointers are welcome. I’m quite new to Ceph :)
Cheers,
Denis
13 May
13 May
2:18 p.m.
osd df is misleading when using external DB devices, they are always
counted as 100% full there
Paul
--
Paul Emmerich
Looking for help with your Ceph cluster? Contact us at https://croit.io
croit GmbH
Freseniusstr. 31h
81247 München
www.croit.io
Tel: +49 89 1896585 90
On Wed, May 13, 2020 at 11:40 AM Denis Krienbühl <denis(a)href.ch> wrote:
Hi
On one of our Ceph clusters, some OSDs have been marked as full. Since
this is a staging cluster that does not have much data on it, this is
strange.
Looking at the full OSDs through “ceph osd df” I figured out that the
space is mostly used by metadata:
SIZE: 122 GiB
USE: 118 GiB
DATA: 2.4 GiB
META: 116 GiB
We run mimic, and for the affected OSDs we use a db device (nvme) in
addition to the primary device (hdd).
In the logs we see the following errors:
2020-05-12 17:10:26.089 7f183f604700 1 bluefs _allocate failed to
allocate 0x400000 on bdev 1, free 0x0; fallback to bdev 2
2020-05-12 17:10:27.113 7f183f604700 1
bluestore(/var/lib/ceph/osd/ceph-8) _balance_bluefs_freespace gifting
0x180a000000~400000 to bluefs
2020-05-12 17:10:27.153 7f183f604700 1 bluefs add_block_extent bdev 2
0x180a000000~400000
We assume it is an issue with Rocksdb, as the following call will quickly
fix the problem:
ceph daemon osd.8 compact
The question is, why is this happening? I would think that “compact" is
something that runs automatically from time to time, but I’m not sure.
Is it on us to run this regularly?
Any pointers are welcome. I’m quite new to Ceph :)
Cheers,
Denis
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
2:19 p.m.
Hi Daniel,
I had the exact same issue in a (virtual) Luminous cluster without
much data in it. The root cause was that my OSDs were too small (10 GB
only) and the rocksDB also grew until manual compaction. I had
configured the small OSDs intentionally because it was never supposed
to store lots of data. Can you provide some more details like 'ceph
odf df'?
Manual compaction did help, but then I recreated the OSDs with 20 GB
each and the issue didn't occur after that.
Regards,
Eugen
Zitat von Denis Krienbühl <denis(a)href.ch>ch>:
Hi
On one of our Ceph clusters, some OSDs have been marked as full.
Since this is a staging cluster that does not have much data on it,
this is strange.
Looking at the full OSDs through “ceph osd df” I figured out that
the space is mostly used by metadata:
SIZE: 122 GiB
USE: 118 GiB
DATA: 2.4 GiB
META: 116 GiB
We run mimic, and for the affected OSDs we use a db device (nvme) in
addition to the primary device (hdd).
In the logs we see the following errors:
2020-05-12 17:10:26.089 7f183f604700 1 bluefs _allocate failed
to allocate 0x400000 on bdev 1, free 0x0; fallback to bdev 2
2020-05-12 17:10:27.113 7f183f604700 1
bluestore(/var/lib/ceph/osd/ceph-8) _balance_bluefs_freespace
gifting 0x180a000000~400000 to bluefs
2020-05-12 17:10:27.153 7f183f604700 1 bluefs add_block_extent
bdev 2 0x180a000000~400000
We assume it is an issue with Rocksdb, as the following call will
quickly fix the problem:
ceph daemon osd.8 compact
The question is, why is this happening? I would think that “compact"
is something that runs automatically from time to time, but I’m not
sure.
Is it on us to run this regularly?
Any pointers are welcome. I’m quite new to Ceph :)
Cheers,
Denis
_______________________________________________
ceph-users mailing list -- ceph-users(a)ceph.io
To unsubscribe send an email to ceph-users-leave(a)ceph.io
2:31 p.m.
Sure, the db device has a size of 22.5G, the primary deice has 100G.
Here’s the complete ceph osd df output of one of the OSDs experiencing this issue:
ID CLASS WEIGHT REWEIGHT SIZE USE DATA OMAP META AVAIL %USE VAR PGS
14 hdd 0.11960 1.00000 122 GiB 118 GiB 2.4 GiB 0 B 116 GiB 4.2 GiB 96.55 3.06 195
I pasted the full output here, since this might not be so readable in the e-mail:
https://pastebin.com/NWYTHwxh <https://pastebin.com/NWYTHwxh>
The OSD in question has the ID 14.
Let me know if there’s anything else I can provide you with.
Cheers,
Denis
On 13 May 2020, at 11:49, Eugen Block
<eblock(a)nde.ag> wrote:
Hi Daniel,
I had the exact same issue in a (virtual) Luminous cluster without much data in it. The
root cause was that my OSDs were too small (10 GB only) and the rocksDB also grew until
manual compaction. I had configured the small OSDs intentionally because it was never
supposed to store lots of data. Can you provide some more details like 'ceph odf
df'?
Manual compaction did help, but then I recreated the OSDs with 20 GB each and the issue
didn't occur after that.
Regards,
Eugen
Zitat von Denis Krienbühl <denis(a)href.ch>ch>:
Hi
On one of our Ceph clusters, some OSDs have been marked as full. Since this is a staging
cluster that does not have much data on it, this is strange.
Looking at the full OSDs through “ceph osd df” I figured out that the space is mostly
used by metadata:
SIZE: 122 GiB
USE: 118 GiB
DATA: 2.4 GiB
META: 116 GiB
We run mimic, and for the affected OSDs we use a db device (nvme) in addition to the
primary device (hdd).
In the logs we see the following errors:
2020-05-12 17:10:26.089 7f183f604700 1 bluefs _allocate failed to allocate 0x400000
on bdev 1, free 0x0; fallback to bdev 2
2020-05-12 17:10:27.113 7f183f604700 1 bluestore(/var/lib/ceph/osd/ceph-8)
_balance_bluefs_freespace gifting 0x180a000000~400000 to bluefs
2020-05-12 17:10:27.153 7f183f604700 1 bluefs add_block_extent bdev 2
0x180a000000~400000
We assume it is an issue with Rocksdb, as the following call will quickly fix the
problem:
ceph daemon osd.8 compact
The question is, why is this happening? I would think that “compact" is something
that runs automatically from time to time, but I’m not sure.
Is it on us to run this regularly?
Any pointers are welcome. I’m quite new to Ceph :)
Cheers,
Denis
_______________________________________________
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
1464
days inactive
1464
days old
3 comments
3 participants
participants (3)
-
Denis Krienbühl -
Eugen Block -
Paul Emmerich