Hi,
I'm not sure yet whether or not this is made worse by config, however - if
I do something along the lines of:
> seq 100 | xargs -P100 -n1 bash -c 'exec rbd.original showmapped'
I'll end up with at least one of the invocations deadlocked like below.
Doing the same on our v10.2.7 clusters
Hi,
While not necessarily CephFS specific - we somehow seem to manage to
frequently end up with objects that have inconsistent omaps. This seems to
be replication (as anecdotally it's a replica that ends up diverging, and
it's at least a few times something that happened after the osd that held
r fencing image access. It's all about
arbitrating modification,
in support of i.e. object-map).
>
> Thanks.
>
>
>> On 20 Apr 2017, at 6:31 AM, Kjetil Jørgensen <kje...@medallia.com> wrote:
>>
>> Hi,
>>
>> As long as you blacklist the old owner by ip,
Hi,
As long as you blacklist the old owner by ip, you should be fine. Do
note that rbd lock remove implicitly also blacklists unless you also
pass rbd lock remove the --rbd_blacklist_on_break_lock=false option.
(that is I think "ceph osd blacklist add a.b.c.d interval" translates
into
Hi,
rados - Does not shard your object (as far as I know, there may be a
striping API, although it may not do quite what you want)
cephfs - implemented on top of rados - does it's own object sharding (I'm
fuzzy on the details)
rbd - implemented on top of rados - does shard into 2^order sized
Hi,
YMMV, riddled with assumptions (image is image-format=2, has one ext4
filesystem, no partition table, ext4 superblock starts at 0x400 and
probably a whole boatload of other stuff, I don't know when ext4
updates s_wtime
of it's superblock, nor if it's actually the superblock last write or last
Hi,
Depending on how you plan to use the omap - you might also want to avoid a
large number of key/value pairs as well. CephFS got it's directory fragment
size capped due to large omaps being painful to deal with (see:
http://tracker.ceph.com/issues/16164 and
reasoning is - two or more machines failing at the same instant that isn't
caused by switch/power is unlikely enough that we'll happily live with it,
it has so far served us well.
-KJ
On Wed, Mar 22, 2017 at 7:06 PM, Kjetil Jørgensen <kje...@medallia.com>
wrote:
>
> For the most part -
for writes.
On Wed, Mar 22, 2017 at 8:05 AM, Adam Carheden <carhe...@ucar.edu> wrote:
> On Tue, Mar 21, 2017 at 1:54 PM, Kjetil Jørgensen <kje...@medallia.com>
> wrote:
>
> >> c. Reads can continue from the single online OSD even in pgs that
> >>
aving 1
> replica
> > > min_size dictates that IO freezes for those objects until min_size
> is
> > > achieved. http://docs.ceph.com/docs/jewel/rados/operations/pools/#
> set-the-number-of-object-replicas
> > <http://docs.ceph.com/docs/jewel/rados/oper
Hi,
rbd_id.vm-100-disk-1 is only a "meta object", IIRC, it's contents will get
you a "prefix", which then gets you on to rbd_header.,
rbd_header.prefix contains block size, striping, etc. The actual data
bearing objects will be named something like rbd_data.prefix.%-016x.
Example - vm-100-disk-1
Hi,
I want/need cephfs-data-scan scan_links, it's in master, although we're
currently on jewel (10.2.5). Am I better off cherry-picking the relevant
commit onto the jewel branch rather than just using master ?
Cheers,
--
Kjetil Joergensen
SRE, Medallia Inc
Phone: +1 (650)
Hi,
so - we're considering moving omap out to faster media than our rather slow
spinning rust. There's been some discussion around this here:
https://github.com/ceph/ceph/pull/6421
Since this hasn't landed in jewel, or the ceph-disk convenience bits -
we're thinking of "other ways" of doing
Hi
On Fri, Oct 7, 2016 at 6:31 AM, Yan, Zheng <uker...@gmail.com> wrote:
> On Fri, Oct 7, 2016 at 8:20 AM, Kjetil Jørgensen <kje...@medallia.com>
> wrote:
> > And - I just saw another recent thread -
> > http://tracker.ceph.com/issues/17177 - can be an explanation
On Fri, Oct 7, 2016 at 4:46 AM, John Spray <jsp...@redhat.com> wrote:
> On Fri, Oct 7, 2016 at 1:05 AM, Kjetil Jørgensen <kje...@medallia.com>
> wrote:
> > Hi,
> >
> > context (i.e. what we're doing): We're migrating (or trying to) migrate
> off
> >
daemon mds.foo scrub_path ?
-KJ
On Thu, Oct 6, 2016 at 5:05 PM, Kjetil Jørgensen <kje...@medallia.com>
wrote:
> Hi,
>
> context (i.e. what we're doing): We're migrating (or trying to) migrate
> off of an nfs server onto cephfs, for a workload that's best described as
> &qu
Hi,
context (i.e. what we're doing): We're migrating (or trying to) migrate off
of an nfs server onto cephfs, for a workload that's best described as "big
piles" of hardlinks. Essentially, we have a set of "sources":
foo/01/
foo/0b/<0b>
.. and so on
bar/02/..
bar/0c/..
.. and so on
Hi,
http://docs.ceph.com/docs/master/rados/operations/cache-tiering/ states
>
> Note A larger hit_set_count results in more RAM consumed by the ceph-osd
> process.
By how much - what order - kb ? mb ? gb ?
After some spelunking - there's osd_hit_set_max_size, is it fair to make
the following
It sounds slightly similar to what I just experienced.
I had one monitor out of three, which seemed to essentially run one core at
full tilt continuously, and had it's virtual address space allocated at the
point where top started calling it Tb. Requests hitting this monitor did
not get very
19 matches
Mail list logo