[openstack-dev] [Ceilometer] Need help with HBase backend

Thomas Maddox thomas.maddox at RACKSPACE.COM
Mon Aug 12 14:29:07 UTC 2013


Hmmm, that's interesting.

That would effect an all-in-one deployment? It's referencing localhost right now; not distributed. My Thrift server is hbase://127.0.0.1:9090/. Or would that still effect it, because it's a software facilitated localhost reference and I'm doing dev inside of a VM (in the cloud) rather than a hardware host?

I really appreciate your help!

-Thomas

From: Stas Maksimov <maksimov at gmail.com<mailto:maksimov at gmail.com>>
Reply-To: OpenStack Development Mailing List <openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>>
Date: Monday, August 12, 2013 9:17 AM
To: OpenStack Development Mailing List <openstack-dev at lists.openstack.org<mailto:openstack-dev at lists.openstack.org>>
Subject: Re: [openstack-dev] [Ceilometer] Need help with HBase backend

Aha, so here it goes. The problem was not caused by monkey-patching or multithreading issues, it was caused by the DevStack VM losing its connection and getting a new address from the DHCP server. Once I fixed the connection issues, the problem with eventlet disappeared.

Hope this helps,
Stas

On 12 August 2013 14:49, Stas Maksimov <maksimov at gmail.com<mailto:maksimov at gmail.com>> wrote:

Hi Thomas,

I definitely saw this before, iirc it was caused by monkey-patching somewhere else in ceilometer. It was fixed in the end before i submitted hbase implementation.

At this moment unfortunately that's all I can recollect on the subject. I'll get back to you if I have an 'aha' moment on this. Feel free to contact me off-list regarding this hbase driver.

Thanks,
Stas.

Hey team,

I am working on a fix for retrieving the latest metadata on a resource rather than the first with the HBase implementation, and I'm running into some trouble when trying to get my dev environment to work with HBase. It looks like a concurrency issue when it tries to store the metering data. I'm getting the following error in my logs (summary):

013-08-11 18:52:33.980 2445 ERROR ceilometer.collector.dispatcher.database [req-3b3c65c9-1a1b-4b5d-bba5-8224f074b176 None None] Second simultaneous read on fileno 7 detected.  Unless you really know what you're doing, make sure that only one greenthread can read any particular socket.  Consider using a pools.Pool. If you do know what you're doing and want to disable this error, call eventlet.debug.hub_prevent_multiple_readers(False)

Full traceback: http://paste.openstack.org/show/43872/

Has anyone else run into this lovely little problem? It looks like the implementation needs to use happybase.ConnectionPool, unless I'm missing something.

Thanks in advance for help! :)

-Thomas

_______________________________________________
OpenStack-dev mailing list
OpenStack-dev at lists.openstack.org<mailto:OpenStack-dev at lists.openstack.org>
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20130812/a2051d06/attachment.html>


More information about the OpenStack-dev mailing list