Re: [Openstack] Flavor metadata quota doesn't work
On 12/01/2017 08:57 AM, si...@turka.nl wrote: Hi, I have created a flavor with the following metadata: quota:disk_write_bytes_sec='10240' This should limit writing to disk to 10240 bytes (10KB/s). I also tried it with a higher number (100MB/s). Using the flavor I have launched an instance and ran a write speed test. For an unknown reason, the metadata seems to be ingored, since I can write with 500+ MB/s to the disk: [centos@vmthresholdtest ~]$ dd if=/dev/zero of=file.bin bs=100M count=15 conv=fdatasync 15+0 records in 15+0 records out 1572864000 bytes (1,6 GB) copied, 2,78904 s, 564 MB/s [centos@vmthresholdtest ~]$ Running Newton. Yeah, that functionality doesn't work. Really, not sure if it ever did. Best, -jay ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] [xenserver][ocata] overcommit vcpu
Hi Adhi, Do you mean that you can’t run two VMs each with 8 vCPUs, or do you mean that you are trying to run one VM with more than 8 vCPUs? I believe that the cpu_allocation_ratio means that you can re-use each physical cpu (thread) up to 16 times with different VMs, but each VM is still restricted to at must the number of physical cpu threads you have on the machine. Bob From: Adhi Priharmanto [mailto:adhi@gmail.com] Sent: 29 November 2017 11:38 To: openstack Subject: [Openstack] [xenserver][ocata] overcommit vcpu Hi, I've been running ocata release with xenserver as hypervisor, my question, how to overcommit vCPU of xenserver ? my xenserver have 4 physical core and 8 vcpu. I already add "cpu_allocation_ratio = 16.0" option into my nova.conf at compute nodes, but its still read 8 vCPUs by scheduler. -- Cheers, Adhi Priharmanto about.me/a_dhi +62-812-82121584 ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
[Openstack] Unshelved instance not being assigned to the proper host
I have PCI passthrough enabled for some of my hypervisior hosts. Initial scheduling for flavors that have the passthrough assigned works properly, i.e. instances are assigned to hypervisors with the resources and once the available pool of resources is consumed creation of new instances fails. When instances are shelved, I'm able to create new instances with the pci passthrough flavor and all is good. The problem that I'm running into is that when I try to unshelve an instance that uses the pci passthrough flavor, the wrong hypervisor, i.e. one with no available pci passthrough resources is selected. For example, I have two hypervisors: one with no GPU cards and the other with 4 GeForce 1080 GPU cards. I have a flavor that allocates one GPU per instance and when I try to unshelve an instance, it gets placed on the hypervisor without any GPUs: 2017-12-01 13:55:52.843 2450 WARNING nova.scheduler.host_manager [req-b7876409-1149-4f8a-837d-46f0023fed10 6136e80cc67e4759843a5f95f7a9a98b 7618d5a893f04349b9feb2ca541a6fce - default default] Selected host: overcloud-novacompute-0.localdomain failed to consume from instance. Error: PCI device request [InstancePCIRequest(alias_name='GeForce1080',count=1,is_new=,request_id=,spec=[{dev_type='type-PCI',product_id='1b06',vendor_id='10de'}])] failed: PciDeviceRequestFailed: PCI device request [InstancePCIRequest(alias_name='GeForce1080',count=1,is_new=,request_id=,spec=[{dev_type='type-PCI',product_id='1b06',vendor_id='10de'}])] failed Is there a setting that I'm missing to use PciPassthroughFilter when unshelving instances? [root@overcloud-controller-0 nova]# nova --version9.1.1[root@overcloud-controller-0 nova]# Thanks in advance for any assistance. Bryan___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] Accessing from and to VM instances without using a floating IP
Hello Bernd, thank you for taking time in answering:) Unfortunately one of the problems in my configuration is that L3 is handled directly from ToR switches which do not support NAT, and as far as I understand NAT should happen at L3 router. So it's not really a matter of will , I actually can't do NAT. :( Moreover I feel the question goes a little deeper than the simple use, or not use, of NAT, what I really want to understand is if I, in order to handle software deployment in my project, HAVE to make all VM instances reachable from outside. This bothers me as I can imagine a number of situations where VM need to be reached only from other VM in the tenant but not from outside. What I'm really looking for is some sort of "out of band" access to the VMs that leaverage on the same mechanism used for metadata. Thanks, Andrea 2017-12-01 14:12 GMT+01:00 Bernd Bausch : > I don't know what works for you, and I am not really a practitioner, but here > are a few suggestions. > > - openstack router set --enable-snat for a short window of time. Of course, > that would give access to the entire internet and only limit the time. > - Use egress rules in security groups, or FWaaS, to limit the instance's > internet access > - Set up a second external network that provides the limited access you need > - Apart from the built-in default L3 router, plugins for other routers like > vyatta are available. Perhaps they provide more features than the L3 router. > > I am sure there are other possibilities. > > Bernd > > -Original Message- > From: Andrea Franceschini [mailto:andrea.franceschini...@gmail.com] > Sent: Friday, December 1, 2017 10:48 AM > To: openstack@lists.openstack.org > Subject: [Openstack] Accessing from and to VM instances without using a > floating IP > > Hello All, > > I'm quite new at Openstack and I'm stil trying to figure out how things works > or are supposed to work. > > This is the scenario. > > Let's imagine we've spun a new instance on a network which is not intended > to reach or to be reached from an external network (absence of NAT support > at L3 or for security/design reasons) > > This istance will be given a cloud-init configuration to upgrade the packages > or the O.S. , but due the absence of external connectivity those operations > will fail. > > What I'm wondering is if there's a way to give this instance a limited "out > of band" access to an external http proxy, just to allow the instance to do > regular maintenance or management stuff, like I said, upgrading packages > connect to some management tool (puppet, chef, ansible...). > > Just like the way metadata-proxy works. > > I've successfully set up a nginx reverse proxy with listener in the tenant's > networks namespace to do the task, but I cannot get rid of the "You're doing > it wrong" feeling. :/ > > I mean I feel like I'm missing something important here, otherwise someone > else would have had the same problem, which seems not to be the case, as I > cannot find any web resources that raises the same question. > > Thanks in advance for any suggestion or direction, > > Andrea > > ___ > Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > Post to : openstack@lists.openstack.org > Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack > ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
[Openstack] Flavor metadata quota doesn't work
Hi, I have created a flavor with the following metadata: quota:disk_write_bytes_sec='10240' This should limit writing to disk to 10240 bytes (10KB/s). I also tried it with a higher number (100MB/s). Using the flavor I have launched an instance and ran a write speed test. For an unknown reason, the metadata seems to be ingored, since I can write with 500+ MB/s to the disk: [centos@vmthresholdtest ~]$ dd if=/dev/zero of=file.bin bs=100M count=15 conv=fdatasync 15+0 records in 15+0 records out 1572864000 bytes (1,6 GB) copied, 2,78904 s, 564 MB/s [centos@vmthresholdtest ~]$ Running Newton. Thanks! Sinan ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] Accessing from and to VM instances without using a floating IP
I don't know what works for you, and I am not really a practitioner, but here are a few suggestions. - openstack router set --enable-snat for a short window of time. Of course, that would give access to the entire internet and only limit the time. - Use egress rules in security groups, or FWaaS, to limit the instance's internet access - Set up a second external network that provides the limited access you need - Apart from the built-in default L3 router, plugins for other routers like vyatta are available. Perhaps they provide more features than the L3 router. I am sure there are other possibilities. Bernd -Original Message- From: Andrea Franceschini [mailto:andrea.franceschini...@gmail.com] Sent: Friday, December 1, 2017 10:48 AM To: openstack@lists.openstack.org Subject: [Openstack] Accessing from and to VM instances without using a floating IP Hello All, I'm quite new at Openstack and I'm stil trying to figure out how things works or are supposed to work. This is the scenario. Let's imagine we've spun a new instance on a network which is not intended to reach or to be reached from an external network (absence of NAT support at L3 or for security/design reasons) This istance will be given a cloud-init configuration to upgrade the packages or the O.S. , but due the absence of external connectivity those operations will fail. What I'm wondering is if there's a way to give this instance a limited "out of band" access to an external http proxy, just to allow the instance to do regular maintenance or management stuff, like I said, upgrading packages connect to some management tool (puppet, chef, ansible...). Just like the way metadata-proxy works. I've successfully set up a nginx reverse proxy with listener in the tenant's networks namespace to do the task, but I cannot get rid of the "You're doing it wrong" feeling. :/ I mean I feel like I'm missing something important here, otherwise someone else would have had the same problem, which seems not to be the case, as I cannot find any web resources that raises the same question. Thanks in advance for any suggestion or direction, Andrea ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
[Openstack] Accessing from and to VM instances without using a floating IP
Hello All, I'm quite new at Openstack and I'm stil trying to figure out how things works or are supposed to work. This is the scenario. Let's imagine we've spun a new instance on a network which is not intended to reach or to be reached from an external network (absence of NAT support at L3 or for security/design reasons) This istance will be given a cloud-init configuration to upgrade the packages or the O.S. , but due the absence of external connectivity those operations will fail. What I'm wondering is if there's a way to give this instance a limited "out of band" access to an external http proxy, just to allow the instance to do regular maintenance or management stuff, like I said, upgrading packages connect to some management tool (puppet, chef, ansible...). Just like the way metadata-proxy works. I've successfully set up a nginx reverse proxy with listener in the tenant's networks namespace to do the task, but I cannot get rid of the "You're doing it wrong" feeling. :/ I mean I feel like I'm missing something important here, otherwise someone else would have had the same problem, which seems not to be the case, as I cannot find any web resources that raises the same question. Thanks in advance for any suggestion or direction, Andrea ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Re: [Openstack] Computer Science Master's Thesis Topics?
Hi Felix! I'm a PhD student at CTU FEE. My own topic was simulation of automatic scaling using both the event-driven approach and queue network modelling. The second part of the thesis was on the prediction of web server workload in time. If you wanted, you could continue and simulate some predictive autoscaling algorithms. The framework is already there, and I have tons of literature on how to continue, but no time, as I'm working in the public cloud industry now. It would also be interesting to extend the prediction, which relies on daily seasonality and do the other approach - short-term anomaly detection. One could also apply performance engineering models to machine learning, to better estimate the length of big data processing jobs. I also have some free practical systems engineering topics from clouds and big data, such as: Compare Hadoop access methods - jump box vs. Apache Knox Automatic server configuration integrated with network management Compare an enterprise disk array with software defined storage Install and benchmark alternate network layers in OpenStack That's about all I have ready at the time. Tomas -Original Message- From: Felix Fischer [mailto:m...@felixfischer.org] Sent: Thursday, November 30, 2017 9:02 PM To: openstack@lists.openstack.org Subject: [Openstack] Computer Science Master's Thesis Topics? Hey Folks, I'm Felix and I'm looking for a topic for my master's thesis in computer science. I studied at the Humboldt University of Berlin and currently I'm working as an OpenStack consultant. I was hoping some of you could point me to areas or even specific topics related to OpenStack which need further scientific research. My topic needs to be about a scientific research question and cannot be a programming task or something similar to that. Maybe some of you have a problem for me which needs solving or can point me directions where to look at or who to ask? Thank you in advance and kind regards, Felix Fischer ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack ___ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : openstack@lists.openstack.org Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack