发起问题

Babar Ismail 的档案 - activity

2020-04-24 05:48:06 -0500 获得奖牌  著名问题 (source)
2020-04-24 05:48:06 -0500 获得奖牌  热门的问题 (source)
2020-04-24 05:48:06 -0500 获得奖牌  受欢迎的问题 (source)
2020-04-01 21:09:51 -0500 问了问题 how is local_gb set?

How does openstack calculate local_gb? There is more than enough space available on my host but the local_gb is calculated to 98GB which severely restricts the number of vms i can host on it. Is there a config property that controls it or is it calculated automatically? If automatic, can anyone give me any pointers to what should i be looking to increase?

2020-04-01 21:09:51 -0500 评论问题 Nova/Newton: wrong free disk reporting?

were you able to figure out the root cause? I am having a similar problem

2020-04-01 18:17:49 -0500 获得奖牌  著名问题 (source)
2019-12-30 22:10:00 -0500 获得奖牌  受欢迎的问题 (source)
2019-12-30 22:10:00 -0500 获得奖牌  热门的问题 (source)
2019-09-14 07:50:55 -0500 获得奖牌  热门的问题 (source)
2019-06-13 15:40:50 -0500 评论问题 where do I specify etcd cluster hosts in neutron?

shouldn't neutron server on the controller node also require etcd configuration?

2019-06-13 01:01:05 -0500 获得奖牌  受欢迎的问题 (source)
2019-06-12 17:10:32 -0500 问了问题 where do I specify etcd cluster hosts in neutron?

In our openstack configuration, etcd cluster is running on the controller nodes and we are using calico for vm networking. If etcd node goes down, we expect openstack to be still able to configure the network as there are other etcd nodes that we can talk to. However, I can't find any configuration where I can specify the etcd nodes. I have a feeling that it is defaulting to localhost. Can someone point me to where i can specify etcd nodes for neutron?

2019-05-08 18:28:39 -0500 获得奖牌  粉丝
2019-05-07 20:42:40 -0500 评论问题 save vm state before reboot

thank you! I'll try that.

2019-05-07 18:23:22 -0500 问了问题 save vm state before reboot

Our case is that if the hypervisors are rebooted, we want the VM's hosted on that hypervisor to come back in the same state they were initially. E.g. reopen all open windows like terminal, firefox, etc. I understand if there are any ssh connections they will be lost but I would still want all the terminal windows to open up.

resume_guests_state_on_host_boot only takes the vm back into ACTIVE state but it doesn't restore the state of the vm before reboot.

Is there any configuration option that will do this? Or any suggestions that we might be able to code up?

2019-02-07 20:34:09 -0500 评论问题 how to configure keystone timeout

We are using an nginx load balancer that has a timeout already configured and controlled by another team. we are unable to change it as it is not under our control. I was hoping if there is a way to control it from keystone.

2019-02-06 16:26:17 -0500 问了问题 how to configure keystone timeout

I have a high availability setup and have 2 keystone hosts behind a vip that has a timeout of 30 seconds. The problem is that it can take longer than 30 seconds for keystone to service a request so I want to configure it stay below 30 seconds.

Is there some config value in keystone.conf I can set that will timeout after 25 seconds?

2019-02-05 00:19:11 -0500 评论问题 how to debug keystone performance?

Is there a separate setting for it? What I read was all resources are set to be cache enabled and as long as the global cache switch is enabled. Can you point me to a config setting which I must ensure is set?

2019-02-04 18:29:29 -0500 问了问题 how to debug keystone performance?

we have an openstack deployment running just fine but a simple curl command like this one: "time curl -k -v --max-time 30 https://localhost:5000%22 (https://localhost:5000") can take up to 30 seconds. For some calls it is very fast (like sub second) but some calls timeout. As far as I understand, it should just be fetching the version from the database. Other services on the box are running just fine.

A load balancer is running that consistently sends this request to the machines. I have toggled the global caching switch to on using memcache backend but that doesn't affect the performance of this simple curl command.