Ask Your Question

Zollner Robert's profile - activity

2016-04-13 05:16:04 -0500 commented answer Instance log

I don't think i can compare them,with StackTach (which is pretty lightweight),I can easily dig trough Instance related events,find out how much time was spend executing each,etc.While Logstash would be better used to collect project wide logs as a way to monitor your complete infrastructure.

2016-04-02 07:14:12 -0500 answered a question Instance log

Last time I checked (about a year ago) the best way was to listen on various notifications OpenStack was publishing to RabbitMQ. From here you could parse and save relevant events to mysql,logstash,whatever.

I have used StackTach (v2) for this, it is a lightweight python app which did the job OK. StackTach v3 seemed to complex back then, an overkill for what i was trying to do.

2014-12-19 09:58:04 -0500 commented question load balcer driver is not loaded

Is neutron starting without lbaas? These errors seem unrelated to your lbaas config.

2014-12-17 03:56:22 -0500 received badge  Enthusiast
2014-12-16 18:23:59 -0500 commented question Howto use VLAN tagged external networks?

The Rackspace multi-part networking blog post that Oreilly pointed to, is probably the best source available. It is very helpful if you find the neutron vocabulary confusing and you have a networking background.

2014-12-16 13:31:55 -0500 received badge  Necromancer (source)
2014-12-16 07:13:50 -0500 answered a question Quantum, DNSMasq, and resolv.conf

At least on havana and icehouse , dhcp neutron agent is preventing dnsmasq to use nameservers from /etc/resolv.conf (using the --no-resolv parameter)

Default behaivior of the dhcp agent is to advertise itself as the dns server to the virtual machines. If you get 8.8.8.8 dns trough dhcp, it was most like manually configured somewhere.

1) You can override this using horizon or cli for every subnet:

neutron subnet-(update|create) --dns-nameservers list=true 192.168.0.254 192.168.1.254 ...

2) Or you can specify global dnsmasq options in the file defined with "dnsmasq_config_file" parameter of the dhcp agent config.

dhcp-option=6, x.x.x.x, y.y.y.y

Upstream DNS forwarders are defined in the dhcp agent config with dnsmasq_dns_servers=8.8.8.8,4.4.4.4 parameter

2014-12-15 22:57:18 -0500 received badge  Teacher (source)
2014-12-15 22:57:18 -0500 received badge  Necromancer (source)
2014-12-15 21:10:58 -0500 answered a question lbaas health monitoring implementation

Health check depends on what lbaas driver you use.

If you use haproxy then neutron lbaas agent will create the haproxy config files with the health_monitor checks you have associated with that LB Pool and then spawn a haproxy process.

The haproxy process will be responsible for the health checks

Some info about lbaas setup : https://openstack.redhat.com/LBaaS

2014-12-15 06:51:51 -0500 answered a question Virtual machine as gateway

1) VM-01 should have two interface one in Netw-A(eth0) and another in Netw-B(eth1) at this point if your "Security Group" setting permits it you should be able to ping between ip addreses of VM-01(eth1) and VM-02 that are in Netw-B

2) Next to be able to ping from Vm-02 to Vm-01 eth0 ipaddr (Netw-B) you must delete some iptables rules or disable the neutron firewall completely

get port id of VM-01 eth1:

neutron port-list |grep "192.168.1.2"
| **6cd7f3ab**-.. fa:16:3e:9c:38:23 | {"subnet_id": "b9...", "ip_address": "192.168.1.2"}|

id = 6cd7f3ab

iptables -n -L -t filter -v  --line-numbers | grep -i "Chain neutron-openvswi-s"6cd7f3ab -A 3
Chain neutron-openvswi-s6cd7f3ab-4 (1 references)
num   pkts bytes target     prot opt in     out   source        destination         
1     2658  229K RETURN     all  --  *      *     192.168.1.2   0.0.0.0/0    MAC FA:16:3e:9c:38:23
2        5     0 DROP       all  --  *      *     0.0.0.0/0     0.0.0.0/0

Delete rule nr. #2:

iptables -D neutron-openvswi-s6cd7f3ab-b 2

3) enable nat on vm-01:

iptables -t nat -A POSTROUTING -o **eth0** -j MASQUERADE