Ask Your Question

jgriffith's profile - activity

2016-12-06 06:12:48 -0500 received badge  Good Answer (source)
2015-08-12 19:38:06 -0500 commented answer Unable to create instances

No, the problem here (at least in my case): I can run a "nova net-list" and the private network is returned. Also, I only run into this error when I try and run the client from a remote machine. If I log on to the Nova node everything works fine. Also, all other nova cmds seem to work ok.

2015-08-03 10:36:16 -0500 commented question Cinder volumes all of the sudden go read-only.

Hi Glen, Are you using the LVM driver? Might try looking in the n-cpu and KVM logs on the system. Sounds like the Instance/KVM is detecting something here and setting the state independent of Cinder. Looking on my side to try and repro.

2015-02-05 11:48:39 -0500 answered a question Volume-attach fails - ERROR No initiators or wwpns supplied.

Your missing info in your Fibre Channel configuration for the IBM Storwize driver, or the Storwize FC driver is flat out busted (I suspect it's a config issue). You will likely have better luck working with IBM support but if you don't have any luck I'll see if I can find a driver maintainer from IBM to help out.

Ahh... so I didn't realize, not only is the FC and StorWize, but also ESX as your hypervisor. So all I can think of is to ensure FC connectivity between your ESX nodes and your FC storage backend.

I'll downvote this on my own now that there are more details here.

2014-12-03 11:34:20 -0500 answered a question How to calculate IOPS for Cinder Storage ?

The answer to this is completely dependent on your deployment and what your users choose to do. There's no set "you need 'n' IOPS for a Cinder backend device. It all comes down to usage, what your users will be doing with the Volumes and what you need.

2014-06-25 23:28:36 -0500 answered a question Creating cinder volume from image is extremely slow

Hi Chris,

Despite some of the other answers you've received elsewhere... yes, this is unfortunately "expected' behavior. The problem is the number of things that have to be done to fetch, convert and load the image. The process goes something like this: 1. Fetch/Download the image from Glances backing store to your Cinder node 2. Run qemu-convert on it to make sure it's valid and that it's raw format suitable for being written straight to a volume 3. Copy the converted image off to the Cinder Volume

There's a bunch of overhead here and it's something we're aware of and know needs fixing. One of the problems is that qemu-convert can't pipe the conversion process out to the device, that alone would help quite a bit.

One thing that I recommend to people is limit this by building template volumes. This way you "suffer" through creating the bootable volume once per image; then just use Cinders clone and extend features to build new bootable volumes off of your templates rather than fetching and converting every time.

Meanwhile, hopefully we'll find some ways to make this better/faster.

2014-06-16 20:31:19 -0500 received badge  Nice Answer (source)
2014-03-11 17:50:17 -0500 received badge  Necromancer (source)
2014-02-20 12:14:28 -0500 answered a question web server on openstack instance

There's some good info here: http://docs.openstack.org/trunk/opens...

You most likely can just add the pre-defined https or https security rules and be good to go. It works for me at least.

2014-02-11 17:40:02 -0500 received badge  Enthusiast
2014-02-09 01:07:30 -0500 answered a question cinder-mange sync is not creating tables in havana

Mili is likely correct, here. Cinder has "found" a database, configured it and is somehow using it, otherwise you would not be able to create volumes. cinder-manage db sync is going to the same mystery DB as well.

2014-02-09 01:05:15 -0500 answered a question How to list iscsi target details like iqn number using cinder client?

No, Cinder does not provide that via the API or client. It's intent is to abstract that out so you don't have to know about it.

That being said it would be trivial to write your own extension to do something along the lines of what you want. Or even easier just use the tgt-admin commands like you mentioned, it's the same info that's stored in the provider_location field.

2014-02-09 01:01:47 -0500 commented answer iscsiadm fails to login to iSER target because of misspelled parameter in the TGT target configuration

Are you using OpenStack community packages or are you using Mellanox's fork found here: https://github.com/mellanox-openstack/cinder-iser? If you're using the Mellanox fork you'd be best to contact them for support.

2014-02-09 00:53:01 -0500 answered a question Can we leverage cinder APIs for FC HBA level or does it only make sense at the FC target level?

I'm not exactly sure what you have in mind here, but currently no. The only thing we offer currently is some basic support for FC targets and that has some caveats. There is work in progress to widen the scope of FC managment and add things like zoning etc.

I'm not sure what you have in mind when you say "FC HBA level" but Cinder's focus is on offering the block storage as a service as opposed to things like protocol management.

2014-02-09 00:50:57 -0500 commented question Cinder Error: arguments already parsed: cannot register CLI option

Can you post your cinder.conf file? Looks like a bad config.

2014-02-09 00:48:57 -0500 answered a question Steps to publish new Cinder volume driver

Hi, The process for submitting a driver isn't well documented, I'll fix that this week and write up some docs on how to get started.

Short summary here: First make sure you go through the "how to contribute to openstack found here: [1] From there you'll need to create a blueprint, here's an example of a driver blueprint: [2] Recommendations other than that, get on IRC and join the #openstack-dev channel, there are always devs there that can help answer questions and point you in the right direction. Make a note of the features that have to be implemented, the best way to start typically is look at the base driver (cinder/volume/driver.py), that will give you all of the implemented methods. Then look at cinder/volume/drivers/lvm.py which is the reference implementation and will show how the base driver is over-ridden, the connector is used etc. ** note the blueprint is an old example but is more detailed than some current versions and has a specs page, browse blueprints.launchpad.net/cinder for others.

update: Added a link to a "how to contribute a driver" to the plugins section of the main Cinder wiki page. https://wiki.openstack.org/wiki/Cinder

2014-02-09 00:37:30 -0500 commented question rdo Hypervisor Summary Disk Usage Used 0 of 49.0GB

Correct, instances do not count against the volume quota (or the "volume count" and "volume summary") numbers shown in the dashboards summary view. I'm not exactly clear what you're seeing, it sounds like you're stating that the dashboard summary shows "0 of 49GB" used? Not sure what this has to do with the hypervisor, or the comment about "change to 40". You list cinder quota-usage for the admin tenant, are you viewing summary in the dashboard as admin tenant as well?

2014-02-07 12:42:42 -0500 received badge  Editor (source)
2014-02-07 12:36:27 -0500 answered a question Cinder Vol error when creating volumes larger than 200G

So after chatting in IRC, this is an issue with your SAN device that you're using. The kernel logs included the message stating you were out of space on the device and couldn't allocate as much as you requested.

Given that you're running an external attached device and laying the Cinder LVM driver over top of it and that the kernel is not agreeing with what you see as "free-space" I don't think this is a Cinder problem at all but more of an error with your backend device, it's reporting and your configuration of the Cinder Volume Group on top of it.

I'd suggest removing cinder from the equation. Since you've attached the SAN Raid device and created the VG, just try lvcreate manually and see if you have the same issue.

2014-02-07 12:27:41 -0500 commented question Cinder Vol error when creating volumes larger than 200G

Important information missing here; this is a SAN attached Raid with LVM on top of it. Also the missing part of the trace: 2014-02-07 11:12:23.586 2022 TRACE cinder.openstack.common.rpc.amqp Stderr: ' device-mapper: resume ioctl failed: Invalid argument\n Unable to resume cinder--volumes1-volume--a0297a8d--0a0e--4244--bc64--d882454fa0ce (252:3)\n Internal error: Maps lock 14544896 < unlock 14548992\n Aborting. Failed to activate new LV to wipe the start of it.\n' Suspect your backend device isn't matching actual capacity to what you setup in the VG. Regardless you should check the system logs for info about your backend device here.

2014-01-13 10:24:13 -0500 answered a question cinder LVMISCSIDriver Unable to update stats, driver is uninitialized

You have an awful lot going on here, it sounds like you've addressed the issue of the volume driver not being initialized, so I'll jump to the delete issues your describing.

Cinder provides a volume_clear option which by default will dd zeros across the entire lv when you delete it. This unfortunately can take an extremely long time. If you don't need this security, you can set the following in your cinder.conf file to skip it:

volume_clear = none

2014-01-09 15:34:29 -0500 commented question cannot create volumes, stuck in creation

Missing some info here... would be good to know: 1. output of "sudo vgs" (did you create your cinder-volumes1 VG) 2. output of "sudo lvs" (did the volume get created from the LVM side) The calls in the API look ok, so I would expect to see something about them in the sched log, error/warn or trace. <update> Any chance of getting the cinder-scheduler logs and cinder-volume logs? One of those two will have some error or trace information related to the problem you're seeing and would likely provide some answers.

2013-12-27 11:11:54 -0500 commented answer cannot remove volume

You need to use the reset-state api method. Change the state from error_deleteing to error so that it's valid for delete. Still need to figure out what's up with the RBD driver that got you here in the first place though.

2013-11-01 13:50:53 -0500 answered a question how can i get volume name in create snapshot call

The create_snapshot method takes a source_volume reference object as a parameter. That volume ref object has the name (and other needed items). Have a look at cinder.volume.drivers.lvm.py for an example.

2013-10-31 22:56:28 -0500 answered a question How can I merge an existing cinder volume to another?

In Grizzly there's really no built in way to do what you want here. You'd have to hack on the code and modify the DB manually. It is possible, but involves manual hacking etc.

2013-10-31 22:52:15 -0500 commented question Creating bootable Cinder volume using volume-type

volume-type and image-id are fine to use together, didn't see your cmd in the posting, should be something like "cinder create --image-id xxx --volume-type xxx $size". Let me know the cmd you're using. Also take a look at the scheduler and volume logs, see what they're reporting.

2013-10-31 22:52:15 -0500 received badge  Commentator
2013-10-31 22:48:03 -0500 commented question How do I change from 1ge to 10ge networking after deployment?

Which backend driver are you using? You should be able to change the iscsi target settings in cinder.conf assuming your compute nodes have network access already configured/setup over your new 10G network.

2013-08-15 00:26:20 -0500 answered a question can cinder be used as an independent storage as a service?

Not really, there's an awful lot of additional plumbing that's involved and while you could hack together some code to do what you suggest (bypass keystone/security, and talk directly to the client to create and get IQN's etc) I'm not sure it would be worth it.

That being said Cinder is a BlockStorageAsAService project and there has been interest in having it used by other cloud projects that this sort of investment and extra code might be worthwhile.

2013-08-15 00:20:22 -0500 commented answer Only the first backend used can create volumes

Not sure but I believe that was a bug in the initial Grizzly release I'd be curious what version you were running there.

2013-08-15 00:15:32 -0500 commented answer For cinder iscsi volumes, is the iscsi stack needed on the VM instances?

actually tgt is the back-end device, the compute node is running iscsi-initiator. But in terms of it running only on the compute node you're correct. It's attached to the compute node and passed in to the instance via libvirt.

2013-08-15 00:10:50 -0500 answered a question cinder consumption

Actually, you're statement was correct, and so was the reponse added by the user above. There are two things that control the amount of block-storage that a tenant/user can deploy.

The first is as you guessed, the available space in the cinder-volume VG. If you're not using thin provisioning then the 500Gb capacity of the VG is in fact all you have for both Volumes and Snapshots. In addition, as the other poster pointed out there are also quotas assigned to each user. This included GB's of volumes and snapshots allowed as well as total number of volumes and snapshots allowed. There are additional quotas as well but those two are the primary ones regarding your question.

As far as where the VG cinder-volumes is located. It's not actually on the controller, it's on what's called a Cinder volume node. The machine/node that is running the cinder volume-service. You can add cinder-volume nodes/services to grow and scale over time. You can also have multiple cinder volume services running on a single machine that points to different VG's or external SAN based storage.

Hope that helps.

2013-06-21 05:20:00 -0500 received badge  Nice Answer (source)
2013-06-18 17:45:41 -0500 answered a question attach volume get exception when instance_uuid is Invalid,but the status of volume is still attaching

You're correct, this is a bug. We should either be checking the UUID prior to the db update as you suggest, or do an update on the status when the uuid_like error is encountered. I've filed a bug [1], thanks!!

[1] https://bugs.launchpad.net/cinder/+bug/1192370

2013-06-18 17:33:36 -0500 answered a question Cinder: Volume Creation Error - iscsi

Change content of the /etc/tgt/targets.conf from "include /etc/tgt/conf.d/*.conf" to: include /etc/tgt/conf.d/cinder_tgt.conf:

include /etc/tgt/conf.d/cinder_tgt.conf
include /etc/tgt/conf.d/cinder.conf
default-driver iscsi

Then restart tgt and cinder-* services so they pick up the new configuration.

I think that might be what you need here.

2013-06-18 17:29:58 -0500 answered a question Cinder: Volumes are getting created with ERROR status

Hi Atreyee,

It seems you've missed the step on creating the cinder-volumes VG on your cinder node. The install guide (http://docs.openstack.org/grizzly/openstack-block-storage/admin/content/cinder-install.html) has a section showing you how to do this with a loop-device for testing which you can do but what you want is to actually do a vgcreate and name is cinder-volumes.

2013-05-10 16:08:15 -0500 answered a question Any hint for a compute Node diskless setup? (only iSCSI NetApp persistant disk for VMs)

Your correct, this feature is not available via the NetApp driver. Here's a wiki page that shows some of the different features and implementations for each driver: https://wiki.openstack.org/wiki/CinderSupportMatrix

2013-05-02 01:23:37 -0500 received badge  Teacher (source)
2013-05-01 17:42:30 -0500 answered a question Cannot reach floating ip from VM

Can you ping your instances via floating IP's outside of OpenStack? I'm wondering if maybe you didn't set the network rules to allow ping and ssh:

http://docs.openstack.org/trunk/openstack-compute/admin/content/enabling-ping-and-ssh-on-vms.html

2013-05-01 17:34:04 -0500 commented question Launching instance and volume

You're not by chance running on a 32 bit machine are you? I would go for dtroyer's answer however it doesn't explain the issue you're having with instance creation, 32 bit however may very well explain both.

2013-04-30 09:34:23 -0500 received badge  Supporter (source)
2013-04-30 09:33:58 -0500 commented answer What are the access credentials for the ubuntu image?

henrik16, in order to get external access you need to enable access, check out: http://docs.openstack.org/trunk/openstack-compute/admin/content/enabling-ping-and-ssh-on-vms.html

2013-04-30 09:30:43 -0500 commented answer Why do I get "No portal found" error while attaching cinder volume to VM?

FYI, posting an answer with a "me too" doesn't help much and adds confusion for folks (like me :) ) when I confuse your config files with the original posters. :) You can add a comment stating that you're experiencing the problem as well and ask for help that way, but it's not an "answer"