Re: [Openstack] [Nova] CPU Scalling, Quota for Disk I/O
Hello, any thoughts on this ? Thank you -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com On Mon, Jul 1, 2013 at 4:22 PM, Bruno Oliveira ~lychinus brunnop.olive...@gmail.com wrote: Hello Stackers, Today morning I saw an interesting question regarding CPU Scaling in the list, which got me to ask the following: Currently (or in the roadmap) do we have any feature on Nova (regardless of the hypervisor underneath) to set maximum disk I/O throughput a VM can have ? I mean, let's say we have hundreds of VMs under the same host as in production, and for some reason we're lacking performance due to one (or a few of them) being too hungry/greedy for disk reads/writes. Question 1) Is there a way we can set quotas for disk I/O for a (group of) instances ? Like: for this one (or this group), don't exceed the threshold of 50 MB/seg Question 2) Also, do we have anything like vertical scalling ? I mean, like defining CPU and Memory Balloons as extra resources that a set of VMs can make use of (temporarily), if they're demanding to ? Note: I've seen some of the videos of Heat talking about increasing horizontally the number of instances behind a load balancer to attend an increasing number of user requests, for instances... Thank you so much. -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] [Nova] CPU Scalling, Quota for Disk I/O
Hello Stackers, Today morning I saw an interesting question regarding CPU Scaling in the list, which got me to ask the following: Currently (or in the roadmap) do we have any feature on Nova (regardless of the hypervisor underneath) to set maximum disk I/O throughput a VM can have ? I mean, let's say we have hundreds of VMs under the same host as in production, and for some reason we're lacking performance due to one (or a few of them) being too hungry/greedy for disk reads/writes. Question 1) Is there a way we can set quotas for disk I/O for a (group of) instances ? Like: for this one (or this group), don't exceed the threshold of 50 MB/seg Question 2) Also, do we have anything like vertical scalling ? I mean, like defining CPU and Memory Balloons as extra resources that a set of VMs can make use of (temporarily), if they're demanding to ? Note: I've seen some of the videos of Heat talking about increasing horizontally the number of instances behind a load balancer to attend an increasing number of user requests, for instances... Thank you so much. -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] [Cinder][HyperV][KVM] iSCSI / NFS/ FCoE for Block Storage Implementation
Dear stackers, salute o As I'm moving forward to finally deploy Openstack to production, I'd like to hear your thoughts for when dealing with Cinder in an approach of Block (iSCSI) vs Filesystem (NFS) As far as I've read, iSCSI is extremely resilient and more reliable than NFS since it already address issues like network faults, by using multiple channels as individual paths to make sure the data reaches its targets. On the other hand, NFS would require the infrastructure itself to guarantee the network connectivity (not that this would be a major issue, though). That for a production use is very important indeed, but I cannot forget of the performance between the two (I've also got to know that NFS might have a superior read IO, due to its read_cache, but it lacks when it comes to writing -- unless I have some sort of write cache, like the one deployed in the ZFS filesystem). Note: we have a Sun/Oracle Storage using the ZFS filesystem for what we're using currently. I've read a lot on the internet, but as you know, I'm not sure how practical these articles can be or if they're just comparing them theorically) **Question 1** Performance-wise, in a very high-throughput network (supposely 10G), would iSCSI perform better than other alternatives ? **Question 2** Would you guys have any thoughts on FCoE over iSCSI ? According to this blueprint, it's already implemented and available for libvirt to use it https://blueprints.launchpad.net/cinder/+spec/fibre-channel-block-storage **Question 3** If NFS is the best option out of the 3, I'm not sure how to deal with it for Hyper-V hosts, I mean, would it be technically possible at all ? **Question 4** [Cloudbase-question] Since the Cinder Volume for Windows Storage Server 2012 doesn't use the libvirt driver, may I ask you if the current build already supports FCoE (Fibre Channel over Ethernet) ? Thank you very very much. Thank you a lot, Stackers. Best regards. -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [Ceilometer][Healthnmon] Dealing with Performance/Monitoring Metrics
@Julien Right. Working on it, Julien. I'll let you guys know if I manage to get any significant progress. Thank you. @Claudio Do we need to install Healthnmon in order to get them? (because ceilometer Yeah. Until we manage to find a way to retrieve'em with Ceilometer or until the two projects merge; -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com On Mon, Jun 24, 2013 at 7:08 AM, claudio marques clau...@onesource.pt wrote: Hi all I think that we are all (people working with ceilometer), trying to figure out how to get the system actual data metrics from VMs. Do we need to install Healthnmon in order to get them? (because ceilometer as Julien already said, can't do it yet) Cheers Cláudio Marques - clau...@onesource.pt http://www.onesource.pt/ From: jul...@danjou.info To: brunnop.olive...@gmail.com Date: Mon, 24 Jun 2013 11:41:13 +0200 CC: openstack@lists.launchpad.net Subject: Re: [Openstack] [Ceilometer][Healthnmon] Dealing with Performance/Monitoring Metrics On Sat, Jun 22 2013, Bruno Oliveira ~lychinus wrote: So, long story short, is technically possible for me to create metrics to make the API show up system_usage_metrics instead of simply system_allocated_metrics ? (I guess I misunderstood, right?) Could you please enlight the path for me ? If you know how to retrieve the actual memory or disk used (rather than allocated) by a VM by measuring from outside the VM, please enlighten us. We (Ceilometer) don't know how to do that, that's why we don't do it actually. -- Julien Danjou ;; Free Software hacker ; freelance consultant ;; http://julien.danjou.info ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] [Ceilometer][Healthnmon] Dealing with Performance/Monitoring Metrics
Dear stackers, please enlight me on something regarding Ceilometer/Healthnmon. So far as I've read and setup, Ceilometer is intended to get metrics for billing usage in larger interval polls. But so far, I've not seen reports of Memory Utilization, Disk Space Utilization or other System Utilization metrics. They're mostly Memory Allocated, Number of vCPUs, and etc. And checking out the Healthnmon projects, I can see that at the very-low level, the mechanism is pretty similiar. I even spotted the following: (...)But in another usage scenario, what if the users want to get the relationship between different measurements? i.e. get all the available measurements related to a specific VM instance, e.g. CPU usage, disk IO usage, network information, storage volumes, …. In Ceilometer it would require extra post-processing of the 'resource_metadata' to find all the measurements. While in Healthnmon, the relationship is already in the data model so it's much easier. The difference between the two data model may comes from that Ceilometer is originally designed for metering, and Healthnmon is designed for monitoring(...) https://wiki.openstack.org/wiki/Ceilometer/CeilometerAndHealthnmon So, long story short, is technically possible for me to create metrics to make the API show up system_usage_metrics instead of simply system_allocated_metrics ? (I guess I misunderstood, right?) Could you please enlight the path for me ? I'm not sure if I got it right... I have seen the blueprints of the projects possibly unifying on Havanna release. But I'm looking for something I can do (for usage metrics) right now (Grizzly). Thank you in advance. -- Bruno Oliveira Developer, Software Engineer irc: lychinus | skype: brunnop.oliveira brunnop.olive...@gmail.com ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [HyperV][Quantum] Quantum dhcp agent not working for Hyper-V
Guys, Any thoughts on this? Like I said, we're not sure if we, either, are doing too much unnecessary things stuff (or just working around it) or if there's a right (and hopefully simpler) way to do it. Any comments or suggestions on this will be very welcome. Thank you. -- On Fri, Jun 7, 2013 at 5:51 PM, Bruno Oliveira ~lychinus brunnop.olive...@gmail.com wrote: (...)Do you have your vSwitch properly configured on your hyper-v host?(...) I can't say for sure, Peter, but I think so... In troubleshooting we did (and are still doing) I can tell that regardless of the network model that we're using (FLAT or VLAN Network), the instance that is provisioned on Hyper-V (for some reason) can't reach the quantum-l3-agent by default (I said default because, we just managed to do it after a hard, long and boring troubleshoting, yet, we're not sure if that's how it should be done, indeed) Since it's not something quick to explain, I'll present the scenario: (I'm not sure if it might be a candidate for a fix in quantum-l3-agent, so quantum-devs might be interested too) Here's how our network interfaces turns out, in our network controller: == External bridge network == Bridge br-eth1 Port br-eth1 Interface br-eth1 type: internal Port eth1.11 Interface eth1.11 Port phy-br-eth1 Interface phy-br-eth1 == Internal network == Bridge br-int Port int-br-eth1 Interface int-br-eth1 Port br-int Interface br-int type: internal Port tapb610a695-46 tag: 1 Interface tapb610a695-46 type: internal Port qr-ef10bef4-fa tag: 1 Interface qr-ef10bef4-fa type: internal == There's another iface named br-ex that we're using for floating_ips, but it has nothing to do with what we're doing right now, so I'm skipping it... So, for the hands-on I know it may be a little bit hard to understand, but I'll do my best trying to explain: 1) the running instance in Hyper-V, which is linked to Hyper-V vSwitch is actually communicating to bridge: br-eth1 (that is in the network controller). NOTE: That's where the DHCP REQUEST (from the instance) lands 2) The interface MAC Address, of that running instance on Hyper-V, is: fa:16:3e:95:95:e4. (we're gonna use it on later steps) Since DHCP is not fully working yet, we had to manually set an IP for that instance: 10.5.5.3 3) From that instance interface, the dhcp_broadcast should be forward - FROM interface eth1.12 TO phy-br-eth1 And FROM interface phy-br-eth1 TO the bridge br-int *** THIS IS WHERE THE PACKETS ARE DROPPED ***. Check it out for the actions:drop - root@osnetwork:~# ovs-dpctl dump-flows br-int |grep 10.5.5.3 in_port(4),eth(src=fa:16:3e:f0:ac:8e,dst=ff:ff:ff:ff:ff:ff),eth_type(0x0806),arp(sip=10.5.5.3,tip=10.5.5.1,op=1,sha=fa:16:3e:f0:ac:8e,tha=00:00:00:00:00:00), packets:20, bytes:1120, used:0.412s, actions:drop - 4) Finally, when the packet reaches the bridge br-int, the DHCP_REQUEST should be forward to the dhcp_interface, that is: tapb610a695-46*** WHICH IS NOT HAPPENING EITHER *** 5) How to fix :: bridge br-eth1 --- 5.1. Getting to know the ifaces of 'br-eth1' --- root@osnetwork:~# ovs-ofctl show br-eth1 OFPT_FEATURES_REPLY (xid=0x1): ver:0x1, dpid:e0db554e164b n_tables:255, n_buffers:256 features: capabilities:0xc7, actions:0xfff 1(eth1.11): addr:e0:db:55:4e:16:4b config: 0 state: 0 current:10GB-FD AUTO_NEG advertised: 1GB-FD 10GB-FD FIBER AUTO_NEG supported: 1GB-FD 10GB-FD FIBER AUTO_NEG 3(phy-br-eth1): addr:26:9b:97:93:b9:70 config: 0 state: 0 current:10GB-FD COPPER LOCAL(br-eth1): addr:e0:db:55:4e:16:4b config: 0 state: 0 OFPT_GET_CONFIG_REPLY (xid=0x3): frags=normal miss_send_len=0 --- 5.2. Adding flow rules to enable passing (instead of dropping) --- # the source mac_address (dl_src) is the from the interface of the # running instance on Hyper-V. This fix the DROP (only) root@osnetwork:~# ovs-ofctl add-flow br-eth1 priority=10,in_port=3,dl_src=fa:16:3e:95:95:e4,actions=normal 6) How to fix :: bridge br-int --- 6.1. Getting to know the ifaces of 'br-int
Re: [Openstack] [HyperV][Ceilometer] Performance statistics from Hyper-V with Ceilometer and libvirt
Alessandro, sure will :) What do I need to get started ? -- * * * Bruno Oliveira* Developer, Software Engineer +55 11 9-6193-3987 skype: brunnop.oliveira brunnop.olive...@gmail.com http://br.linkedin.com/in/brunnopoliveira http://www.twitter.com/lychinus http://www.facebook.com/lychinus http://gplus.to/lychinus On Tue, Jun 11, 2013 at 2:28 PM, Alessandro Pilotti apilo...@cloudbasesolutions.com wrote: Hi Bruno, We just started implementing the Ceilometer Hyper-V inspector for the compute agent (see hyper-v-agent blueprint). Let me know if you'd like to help in testing it. :-) Thanks, Alessandro On Jun 6, 2013, at 00:40 , Bruno Oliveira brunnop.olive...@gmail.com wrote: Dear Stackers, Please I'd like to ask your expertise on ceilometer to try an approach for monitoring a compute nova running hyper-v I got a running environment of devstack with KVM on a single-node machine being fully monitored by ceilometer. I'm being able to access its API to see the data collected as well. On the other hand, I'm using Cloudbase's (cloudbase.it) Openstack Compute Hyper-V Installer (driver), which greatly help me deploying VMs from the devstack node to the hyper-v server. All is working smoothly. So far, in the irc channel #openstack-ceilometer, I got to know that ceilometer, just like collectd, uses libvirt to query the hypervisors for data (thanks dhellmann !). BUT if check in the libvirt.org site, it's being said that there's support for Hyper-V: http://libvirt.org/drvhyperv.html (given a uri to connect to it). 1. So I'm wondering if all I need to get Ceilometer working for windows would be compiling it under my windows envionment with mingw chaintool, for example? 2. Has anyone ever tried this or got any other way to make it data collection from hyper-v successful ? 3. Do you guys have any other approach that you would suggest (installing snmp on each of the cloud servers/instances is not an option) ? Please, share your thoughts. I'd greatly appreciate it. Thank you very very much. -- Bruno Oliveira Developer, Software Engineer ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [HyperV][Quantum] Quantum dhcp agent not working for Hyper-V
(...)Do you have your vSwitch properly configured on your hyper-v host?(...) I can't say for sure, Peter, but I think so... In troubleshooting we did (and are still doing) I can tell that regardless of the network model that we're using (FLAT or VLAN Network), the instance that is provisioned on Hyper-V (for some reason) can't reach the quantum-l3-agent by default (I said default because, we just managed to do it after a hard, long and boring troubleshoting, yet, we're not sure if that's how it should be done, indeed) Since it's not something quick to explain, I'll present the scenario: (I'm not sure if it might be a candidate for a fix in quantum-l3-agent, so quantum-devs might be interested too) Here's how our network interfaces turns out, in our network controller: == External bridge network == Bridge br-eth1 Port br-eth1 Interface br-eth1 type: internal Port eth1.11 Interface eth1.11 Port phy-br-eth1 Interface phy-br-eth1 == Internal network == Bridge br-int Port int-br-eth1 Interface int-br-eth1 Port br-int Interface br-int type: internal Port tapb610a695-46 tag: 1 Interface tapb610a695-46 type: internal Port qr-ef10bef4-fa tag: 1 Interface qr-ef10bef4-fa type: internal == There's another iface named br-ex that we're using for floating_ips, but it has nothing to do with what we're doing right now, so I'm skipping it... So, for the hands-on I know it may be a little bit hard to understand, but I'll do my best trying to explain: 1) the running instance in Hyper-V, which is linked to Hyper-V vSwitch is actually communicating to bridge: br-eth1 (that is in the network controller). NOTE: That's where the DHCP REQUEST (from the instance) lands 2) The interface MAC Address, of that running instance on Hyper-V, is: fa:16:3e:95:95:e4. (we're gonna use it on later steps) Since DHCP is not fully working yet, we had to manually set an IP for that instance: 10.5.5.3 3) From that instance interface, the dhcp_broadcast should be forward - FROM interface eth1.12 TO phy-br-eth1 And FROM interface phy-br-eth1 TO the bridge br-int *** THIS IS WHERE THE PACKETS ARE DROPPED ***. Check it out for the actions:drop - root@osnetwork:~# ovs-dpctl dump-flows br-int |grep 10.5.5.3 in_port(4),eth(src=fa:16:3e:f0:ac:8e,dst=ff:ff:ff:ff:ff:ff),eth_type(0x0806),arp(sip=10.5.5.3,tip=10.5.5.1,op=1,sha=fa:16:3e:f0:ac:8e,tha=00:00:00:00:00:00), packets:20, bytes:1120, used:0.412s, actions:drop - 4) Finally, when the packet reaches the bridge br-int, the DHCP_REQUEST should be forward to the dhcp_interface, that is: tapb610a695-46*** WHICH IS NOT HAPPENING EITHER *** 5) How to fix :: bridge br-eth1 --- 5.1. Getting to know the ifaces of 'br-eth1' --- root@osnetwork:~# ovs-ofctl show br-eth1 OFPT_FEATURES_REPLY (xid=0x1): ver:0x1, dpid:e0db554e164b n_tables:255, n_buffers:256 features: capabilities:0xc7, actions:0xfff 1(eth1.11): addr:e0:db:55:4e:16:4b config: 0 state: 0 current:10GB-FD AUTO_NEG advertised: 1GB-FD 10GB-FD FIBER AUTO_NEG supported: 1GB-FD 10GB-FD FIBER AUTO_NEG 3(phy-br-eth1): addr:26:9b:97:93:b9:70 config: 0 state: 0 current:10GB-FD COPPER LOCAL(br-eth1): addr:e0:db:55:4e:16:4b config: 0 state: 0 OFPT_GET_CONFIG_REPLY (xid=0x3): frags=normal miss_send_len=0 --- 5.2. Adding flow rules to enable passing (instead of dropping) --- # the source mac_address (dl_src) is the from the interface of the # running instance on Hyper-V. This fix the DROP (only) root@osnetwork:~# ovs-ofctl add-flow br-eth1 priority=10,in_port=3,dl_src=fa:16:3e:95:95:e4,actions=normal 6) How to fix :: bridge br-int --- 6.1. Getting to know the ifaces of 'br-int' --- root@osnetwork:~# ovs-ofctl show br-int OFPT_FEATURES_REPLY (xid=0x1): ver:0x1, dpid:92976d64274d n_tables:255, n_buffers:256 features: capabilities:0xc7, actions:0xfff 1(tapb610a695-46): addr:19:01:00:00:00:00 config: PORT_DOWN state: LINK_DOWN 4(int-br-eth1): addr:5a:56:e1:53:e9:90 config: 0 state: 0 current:10GB-FD COPPER 5(qr-ef10bef4-fa): addr:19:01:00:00:00:00 config: