ack.org/admin-guide-cloud/content/section_configuring-compute-migrations.html
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
diculous amount)
So it probably does not fit in 32GB maybe just in 64GB.
So you should probably see quite a lot of cache misses currently in the xfs
stats (xfs xs_dir_lookup, xs_ig_missed)
Cheers,
Robert van Leeuwen
-----Ursprüngliche Nachricht-
Von: Robert van Leeuwen [mailto:robert.v
> I set the vfs_cache_pressure to 10 and moved container- and account-server to
> SSD harddrives.
> The normal performance for object writes and reads are quite ok.
> But why takes moving some partions to only two new harddisks so much time?
> Will it be faster if I add more memory?
My guess: Pr
ve:
We removed all config 1.X nodes to nodes with config 2.0.
We stopped using flashcache, since it was a bit of a pain and without it is
also fine for us.
As a result a full sync when adding an empty new node is now down to 1-2 days
while this used to be 2+ weeks.
is in a specific host aggregate
it will not spawn machines.
However I would certainly want this to be an option.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : ope
age/Tuning_KVM
We saw a major improvement using RAW vs QCOW2
(in theory this should be similar, however, when you need a high random IO
throughput we did see major improvements in our setup with RAW)
On more thing to check is partition alignment. (not very likely for a recent OS)
Ch
omething similar to see if the storage is the actual bottleneck or
if it is something on the client.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openst
>I want to build only private storage cloud using openstack swift.
>So, does i need to install nova compute engine. I have been installed and
>setup swift components.
No, you do not.
You just need Swift and, optionally, keystone.
Cheers,
Robert va
space specific)
fleet: https://coreos.com/using-coreos/clustering/
openshift / openshift origin: https://www.openshift.com/products/origin
Geard (openshift uses this):
https://github.com/openshift/geard/blob/master/README.md
I'd be happy to hear what other people have experience wi
freesoace or so it will
stop), rabbitmq is keep running but will stop processing messages.
Have a look at the rabbitmq management interface:
https://www.rabbitmq.com/management.html
This is usually points you in the right direction.
Cheers,
aproxy ;)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
ck-ported.
The bug report is here: https://bugs.launchpad.net/nova/+bug/1362129
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscri
here is a nice writeup about what is shared when using cells/regions/av
zones/aggregates here:
http://docs.openstack.org/openstack-ops/content/scaling.html
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/
e key set to false +
modify other flavors with this key set to false.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.open
1df-cf: mtu 1500 qdisc
pfifo_fast state UP qlen 1000
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi
ts through e.g. a load-balancer?
Maybe that has some issues with larger requests...
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsu
cehouse RDO here on SL6.
I do not see this error when running this command on our production environment.
As a test I also spawned 100 vm's in dev on a single hypervisor, also no issues.
Running neutron with ML2 plugin and openvswitch.
Cheers,
Robert van Leeuwen
ch some output and configuration
Have you looked if the flows in the ovs are properly created?
If there are no flows you have no traffic.
Do a ovs-ofctl dump-flows on the two bridges to check those..
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.ope
ists migrations will fail.
Maybe this can help fix this bug:
https://bugs.launchpad.net/nova/+bug/1270825
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.opensta
> Apparently, "service neutron-server restart" worked!
> Can I safely ignore the "neuton-db-manage" error?!
In my experience upgrading (from Folsom up to Icehouse) the command should run
without errors or things are broken at some level...
d it works for us..
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
ddress until the DHCP
> process is successful.
I disagree.
I think It is just something you need to be aware of ( I never found this
strange behaviour)
It is hugely useful to know what IP the machine should get, even if the machine
is not (yet) running or suspended/stopped.
Ch
> 1. Assuming layer 7 is what would be used for openstack services like
> keystone and Swift, not layer 4? I saw a layer 4 in front of a layer 7
> in one of the diagrams, didnt really understand why that was needed.
Did not make that drawing ;)
I see no reason why it would be mandatory.
> 2. Do t
and numad (kernel panics) so we
only run it on our hypervisors (which is rock solid).
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe
ms so
I'm not sure if that would be the preferred route though.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lis
hite:
http://engineering.spilgames.com/using-ceilometer-graphite/
Cheers,
Robert van Leeuwen
Sent from my iPad
> On 16 aug. 2014, at 02:14, "Guillermo Alvarado"
> wrote:
>
> Hello,
>
> I am wondering why Ceilometer uses MongoDB instead MySQL or some relational
> dat
over a http based protocol ;)
There are some tools where you can "mount" a swift store as a directory on your
machine so less savvy users can use it to copy some files.
( e.g.
http://www.mirantis.com/openstack-portal/the-comparison-of-openstack-swift-comp
much rather have a "extra_allowed_headers" or something similar that leaves
the defaults for what it is.
> Maybe this will help:
> https://bugs.launchpad.net/swift/+bug/1354312<https://bugs.launchpad.net/swift/+bug/1354312>
Thx for filing a bug.
I've put myself in t
Example: Cache-Control is the one that we added in our setup:
allowed_headers = Content-Disposition, Content-Encoding, X-Delete-At,
X-Object-Manifest, X-Static-Large-Object, Cache-Control
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.opensta
>Where Can i increase the port quota ? Can you please help me with it ?
http://docs.openstack.org/user-guide-admin/content/cli_set_quotas.html#cfg_quotas_common
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mail
014-07-28 13:24:59.389 2797 TRACE nova.compute.manager [instance:
> 4b459e38-ab20-42e5-ae79-1e98a0730c85] PortLimitExceeded: Maximum number of
> ports exceeded
That looks to be the issue.
Increase the port quota global / of the tenant and you should be
to beat.
It might not fit your exact needs today but in a few years there will be very
few reasons to buy a SAN...
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@
lots of
reads when writing.
( I'm running CentOS. )
I'd do some performance testing for your expected workload.
Starting a vm and doing a fio benchmark is not that time intensive :)
Cheers,
Robert van Leeuwen
From: Nhan Cao [nhanc...@gmail.com]
Sen
is necessary
.
I was talking about the debug logs on the server side, not client-side ;)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
also gives a
good indication on whats wrong.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
h group is allowed what.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
/bugs.launchpad.net/nova/+bug/1270825
I'm not sure what the state of migrations is in Icehouse.
There were talks to move the migration tasks to the conductor/schedule(?) in
the codebase but I haven't read anything in the release notes about this.
Seems like there is still some
Hi,
We have some custom rules to have a shared network that only people member of a
specific group could see.
Basically a rule like this:
"shared": "field:networks:shared=True and not field:networks:name=mil",
This no longer seem to work after a havana -> icehouse upgrade.
Everyone sees these net
fixed in the Havana release 2013.2.3.
Cheers,
Robert van Leeuwen
From: Sam Stoelinga [sammiest...@gmail.com]
Sent: Friday, May 09, 2014 10:13 AM
To: openstack@lists.openstack.org
Subject: [Openstack] Boot from Volume stuck at Booting from hard disk
When using
)
* Attaching / detaching cinder volumes (
https://bugs.launchpad.net/nova/+bug/1315354 )
Are these limitations to be expected?
If so, is there any documentation on functionality that is should or should not
work?
Thx,
Robert van Leeuwen
___
Mailing
> What changes will be there in authentication requests for accessing resources
> over many regions ? Will
> access URLs change for specific regions?
The keystone URL will be the same but you will create endpoints in keystone for
each region.
The client will need to set their region in the requ
to have thousands of networks we do not expect to run
out of VLANs)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
ant-id --provider:network_type vlan
--provider:physical_network default --provider:segmentation_id $vlan-id
--shared True
Since this is just a regular VLAN network you can directly patch your router
for the segment also in that VLAN and things should work :)
Cheers,
Robert van Leeuwen
_
space).
You would be able to connect to rabbitmq (so no errors in the logs) but it
stopped processing messages.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.open
f the agent does
not pick it up.
If it never hits the messagebus I guess this is not related to the neutron
agent but the neutron (api) server code or Nova.
Check the Nova and Neutron server logs to see if the message is actually being
put the messagebus.
Could give an understanding where/why this
/ ask for help in
the maillinglists / create a bug report with all the info you can gather.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe
envswitch setup with security groups enabled needs these
devices.
This is because without them traffic will never hit iptables on the compute
node.
There is currently work in progress to copy the iptables functionality to
openvswitch/openflow rules so this setup is no longer needed
older version though.
(e.g. it currently works, there is nothing in the release notes we need, so
lets not break it ;)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : ope
on the compute nodes also need to be the correct
version.
Best would be to have just one version in your local repositorie or use
puppet/chef/whatever ensure => version for this.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.opens
> any tip or experience regarding what happened ?
> thanks as allways
I would start by doing a tcpdump on one of the saturated links on the compute
node to see what kind of traffic it is...
Cheers,
Robert van Leeuwen
___
Mailing list
the new node is properly set to the same id)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
bbit related issues:
https://www.rabbitmq.com/management.html
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack
have pro's and con's.
Cheers,
Robert van Leeuwen
From: Aryeh Friedman [aryeh.fried...@gmail.com]
Sent: Friday, February 14, 2014 7:46 AM
To: openstack@lists.openstack.org
Subject: [Openstack] what is the currently "recommended" host OS for
solution but I could not find
any 3rd party drivers yet.
Drivers I could find seem to be limited to layer 2 networking.
If there is not anything available yet, how much effort would be to create such
a driver?
( days/weeks/months of work ?)
Thx,
Robert van Leeuwen
We upgraded from Grizzly to Havana a while ago using the RDO packages (on SL
6.4)
Using Neutron with VLANs.
What kind of issues are you seeing?
Cheers,
Robert van Leeuwen
From: Rajshree Thorat [rajshree.tho...@gslab.com]
Sent: Monday, December 16
pgrades are a good moment to test your knowledge /
troubleshooting skills on dev ;)
Usually checking what happens in the logfiles gives you enough information on
where to look.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bi
, painless.
(Just had a few small bumps with the db upgrades. Make sure you have a
mysqldump before doing those. :)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@l
not hit it.
We are currently using 3 and will look into 1 & 2 after the Havana release. (if
the l3-agent is still not possible with HA.)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to
you are running the same openvswitch version on those machines
with:
modinfo openvswitch
At a slightly related note:
At some point we decided that GRE was giving us more headaches than that it
solved problems so we moved to a VLAN setup.
Cheers,
Robert van Leeuwen
__
ugh:
quantum router-create router1
quantum router-interface-add router1 ${subnetname)
What is the output of the following?:
quantum l3-agent-list-hosting-router $(routername}
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.
emoving containers)
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
hich is caused by the rsyncs from the
replicators.
With our nodes being quite slow due to the number of files the traffic it is
probably a lot less then you are seeing.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-b
node to be synced.
Usually it goes pretty quickly at first (30 MB/second) and the performance
gradually degrades when the disks start filling up and the machines are running
low on memory.
We have about 6TB on a node to sync.
Cheers,
Robert van Leeuwen
From:
down considerably.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to : openstack@lists.openstack.org
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Hi,
First you need to make sure you create enough containers.
Especially if you the container servers on spinning disks you should keep it
below 1 million objects.
However, I would highly recommend running the containers on SSD's.
Another thing you will have to take in account is the number of
open source but they also have payed/supported version. For a very
quick look you can use their hosted version with a free trial.
Cheers,
Robert van Leeuwen
Sent from my iPad
On 12 aug. 2013, at 18:46, "Steve Heistand" wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> So it seems a big number of inodes considering that we are using just a 15%
> of the total capacity. A different thing here is that we are using 512K of
> inode size and we have a big amount of memory .
> Also we always have one of our disks close to 100% of util, and this is
> caused by the o
e to
significantly improve performance by adding more of those machines and doing
some re-balancing of the load.
Cheers,
Robert van Leeuwen
___
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to :
68 matches
Mail list logo
