Hi Dan,
we are alreay back on the kernel module since the same problems were
happening with fuse. I had no special ulimit settings for the
fuse-process, so that could have been an issue there.
I was pasting you the kernel messages during such incidents here:
http://pastebin.com/X5JRe1v3
I was
Hi,
I'm just writing to you to stress out what others have already said,
because it is very important that you take it very seriously.
On 20/04/15 19:17, J-P Methot wrote:
On 4/20/2015 11:01 AM, Christian Balzer wrote:
This is similar to another thread running right now, but since our
Hi Onur,
actual 50, ideal 330128, fragmentation factor 0.97%
so fragmentation is not an issue here.
Regards,
Christian
Am 20.04.2015 um 16:41 schrieb Onur BEKTAS:
Hi,
Check xfs fregmentation factor for rbd disks i.e.
xfs_db -c frag -r /dev/sdX
if it is high, try defrag
Hi Christian,
Am 13.04.15 um 12:54 schrieb Christian Balzer:
Hello,
On Mon, 13 Apr 2015 11:03:24 +0200 Götz Reinicke - IT Koordinator wrote:
Dear ceph users,
we are planing a ceph storage cluster from scratch. Might be up to 1 PB
within the next 3 years, multiple buildings, new network
So I had extra drives on my lab cluster's admin node and decided to use
them for more OSDs. The weird thing is, unlike all the other nodes, the
OSDs don't start on boot - i have to manually activate them whenever the
cluster is rebooted. Manually running service ceph-all start doesn't start
them
Hi Christian,
I've never debugged the kernel client either, so I don't know how to
increase debugging. (I don't see any useful parms on the kernel
modules).
Your log looks like the client just stops communicating with the ceph
cluster. Is iptables getting in the way ?
Cheers, Dan
On Tue, Apr
Hi Dan,
nope, we have no iptables rules on those hosts and the gateway is on the
same subnet as the ceph cluster.
I will see if I can find some informations on how to debug the rbd
kernel module (any suggestions are appreciated :))
Regards,
Christian
Am 21.04.2015 um 10:20 schrieb Dan van der
Yep, you have hit bug 11429. At some point, you removed a pool and then
restarted these osds. Due to the original bug, 10617, those osds never
actually removed the pgs in that pool. I'm working on a fix, or you can
manually remove pgs corresponding to pools which no longer exist from the
Hello,
On Tue, 21 Apr 2015 08:33:21 +0200 Götz Reinicke - IT Koordinator wrote:
Hi Christian,
Am 13.04.15 um 12:54 schrieb Christian Balzer:
Hello,
On Mon, 13 Apr 2015 11:03:24 +0200 Götz Reinicke - IT Koordinator
wrote:
Dear ceph users,
we are planing a ceph storage
On Tue, Apr 21, 2015 at 3:43 PM, Ilya Dryomov idryo...@gmail.com wrote:
On Tue, Apr 21, 2015 at 11:49 AM, Stefan Priebe - Profihost AG
s.pri...@profihost.ag wrote:
Hi,
while running firefly i've seen that each osd log prints:
2015-04-21 10:24:49.498048 7fa9e925d780 0
Hi Huseyin,
Thanks for responding. I did spend quite a bit of time trying to sort out the
nova and libvirt uid/gid so hopefully they are correct. That is also one of the
reasons why I resorted to testing with a basic text file as root to rule that
out.
Host 2 :
On 04/21/2015 03:04 AM, Andrei Mikhailovsky wrote:
Hi
I have been testing the Samsung 840 Pro (128gb) for quite sometime and I
can also confirm that this drive is unsuitable for osd journal. The
performance and latency that I get from these drives (according to ceph
osd perf) are between 10 -
Thank you everyone for your replies. We are currently in the process of
selecting new drives for journaling to replace the samsung drives. We're
running our own tests using DD and the command found here :
Hi!
I have updated my cluster to Hammer and got a warning too many PGs per OSD
(2240 max 300).
I know, that there is no way to decrease number of page groups, so I want to
re-create my pools with less pg number, move all my data to them, delete old
pools and rename new pools as the old ones.
Dear Neville,
Could you please share output of ls -l /var/lib/nova/instances on both
hosts? The user id's of nova on both is probably different. You can
check /etc/passwd files on both host.
Regards,
Huseyin COTUK
On 21-04-2015 15:43, Neville wrote:
I'm trying to setup live migration in
Hi all,
may there be a problem with the crush function during 'from scratch'
installation of 0.94.1-0 ?
This has been tested many times, with ceph-deploy-1.5.22-0 or
ceph-deploy-1.5.23-0. Platform RHEL7.
Each time, the new cluster ends up in a weird state never seen on my
previous
Hi,
while running firefly i've seen that each osd log prints:
2015-04-21 10:24:49.498048 7fa9e925d780 0
xfsfilestorebackend(/ceph/osd.49/) detect_feature: extsize is disabled
by conf
The firefly release notes V0.80.4:
osd: disable XFS extsize hint by default (#8830, Samuel Just)
But hammer
Hi
I have been testing the Samsung 840 Pro (128gb) for quite sometime and I can
also confirm that this drive is unsuitable for osd journal. The performance and
latency that I get from these drives (according to ceph osd perf) are between
10 - 15 times slower compared to the Intel 520. The
On Tue, 21 Apr 2015, Ray Sun wrote:
Cephers,
Many people told me ceph is ready for production except the cephFS, is this
true? Any why it is? Can any one explain this to me? Thanks a lot.
Ready for production is a subjective judgement call. The reason why we
are cautious when recommending it
On 04/20/2015 04:18 PM, Robert LeBlanc wrote:
You usually won't end up with more than the size number of replicas, even
in a failure situation. Although technically more than size number of OSDs
may have the data (if the OSD comes back in service, the journal may be used
to quickly get
Hi Robert,
I'm running the latest version of Icehouse - 2014.1.4. I know there was some
talk of patching nova but not sure if this was ever done, if it was I assume it
wasn't backported for Icehouse. I'm not specifically worried about needing the
instances folder on shared storage to get it
I think your are using an old version of OpenStack. I seem to remember a
discussion about a patch to remove the requirement of shared storage for
live migration on Ceph RBD. Are you using librbd in open stack?
Robert LeBlanc
Sent from a mobile device please excuse any typos.
On Apr 21, 2015 6:43
I'm trying to setup live migration in Openstack using Ceph RBD backed volumes.
From what I understand I also need to put the libvirt folder
/var/lib/nova/instances on shared storage for it to work as Nova tests for this
as part of the migration process. I decided to look at using CephFS for
On Tue, Apr 21, 2015 at 11:49 AM, Stefan Priebe - Profihost AG
s.pri...@profihost.ag wrote:
Hi,
while running firefly i've seen that each osd log prints:
2015-04-21 10:24:49.498048 7fa9e925d780 0
xfsfilestorebackend(/ceph/osd.49/) detect_feature: extsize is disabled
by conf
The firefly
On 04/21/2015 09:08 AM, Robert LeBlanc wrote:
Your logic isn't quite right and from what I understand, this is how it works:
step choose firstn 2 type rack # Choose two racks from the CRUSH map
(my CRUSH only has two, so select both of them)
step chooseleaf firstn 2 type host #
Thanks for elaborating on those facilities, Greg. Its all starting to make more
sense if I think of it from an osd tree view and the type hierarchy. Figures...
you will eliminate them right around the time that I fully understand how to
use them effectively.
On 04/21/2015 09:52 AM, Gregory
Just want to add my own experience, as I'm using consumer Samsung SSDs
at the moment (Ceph 0.87.1, replication 3, 16 Gbps infiniband).
Originally I only had Samsung 840 EVO 1TB SSDs, which I partitioned with
an initial small partition for the journal and the rest for the OSD
(using XFS). I
Sage,
I have the same question about the road map. I remember in OpenStack
HongKong summit, there was a road map to show ceph will support VMWare in
Q3 last year, but after Redhat acquired Ceph, I think the road map had
changed. Could you please provide more detailed about the road map.
Thanks.
Hi sage,
When can we expect the fully functional fsck for cephfs?. Can we get at
next major release?. Is there any roadmap or time frame for the fully
functional fsck release?
Thanks Regards
K.Mohamed Pakkeer
On 21 Apr 2015 20:57, Sage Weil s...@newdream.net wrote:
On Tue, 21 Apr 2015, Ray
The CRUSH min and max sizes are part of the ruleset facilities that we're
slowly removing because they turned out to have no utility and be overly
complicated to understand. You should probably just set them all to 1 and
10.
The intention behind them was that you could have a single ruleset which
Your logic isn't quite right and from what I understand, this is how it
works:
step choose firstn 2 type rack # Choose two racks from the CRUSH map
(my CRUSH only has two, so select both of them)
step chooseleaf firstn 2 type host # From the set chosen previously (two
racks), select a leaf
Hi,
I've recently seen some confusion over the number of PGs per pool versus
per cluster on the mailing list. I also set too many PGs per pool b/c of this
confusion.
IMO, it is fairly confusing to talk about PGs on the Pool page, but only
vaguely talk about the number of PGs for the
32 matches
Mail list logo
