Re: [CentOS-virt] lvm cache + qemu-kvm stops working after about 20GB of writes
Hello everyone, I've tried just about everything the last weeks and my findings are: - the problem with LVM cache seems NOT to be caused by KVM/qemu. But is /seems/ that it is noticeable more inside a KVM. So the slowdown of the cache also happens on HW-node, but you must give it a serious go before you notice. - Not a single time I succeeded in creating a well working cache using LVM2 cache. In artificial / KVM setups and with *small* devices it works (dm-testsuite etc). But in real life scenario with *fully population all the PV's* on 2TB HDDs and 250G SSDs (both RAID 1) the cache stopped working after 20 - 50 GB of writes although the cache is 150+G large. Please use fio examples below and always use new filenames so not the same blocks are overwritten. The poor performance stayed most of the time even when all blocks were flushed. Very unpredictable cache performance / behavior. I finally decided to go for dm-writeboost in stead of lvm2 cache (dm-cache). This was the only way to create a well working cache that works till 95% filled. But of course would be nicer to have something that generally more stable like LVM2. I guess in the sens of this mailing-list this issue is resolved because it does not seem to belong here. -- Met vriendelijke groet, Richard Landsman http://rimote.nl T: +31 (0)50 - 763 04 07 (ma-vr 9:00 tot 18:00) 24/7 bij storingen: +31 (0)6 - 4388 7949 @RimoteSaS (Twitter Serviceberichten/security updates) On 04/20/2017 04:23 PM, Sandro Bonazzola wrote: On Thu, Apr 20, 2017 at 12:32 PM, Richard Landsman - Rimote <rich...@rimote.nl <mailto:rich...@rimote.nl>> wrote: Hello everyone, Anybody had the chance to test out this setup and reproduce the problem? I assumed it would be something that's used often these days and a solution would benefit a lot of users. If can be of any assistance please contact me. I haven't seen any additional report of this happening, can you please try to reproduce with the new qemu-kvm-ev-2.6.0-28.el7_3.9.1 currently in testing? -- Met vriendelijke groet, Richard Landsman http://rimote.nl T: +31 (0)50 - 763 04 07 (ma-vr 9:00 tot 18:00) 24/7 bij storingen: +31 (0)6 - 4388 7949 @RimoteSaS (Twitter Serviceberichten/security updates) -- SANDRO BONAZZOLA ASSOCIATE MANAGER, SOFTWARE ENGINEERING, EMEA ENG VIRTUALIZATION R Red Hat EMEA <https://www.redhat.com/> <https://red.ht/sig> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> ___ CentOS-virt mailing list CentOS-virt@centos.org https://lists.centos.org/mailman/listinfo/centos-virt ___ CentOS-virt mailing list CentOS-virt@centos.org https://lists.centos.org/mailman/listinfo/centos-virt
Re: [CentOS-virt] lvm cache + qemu-kvm stops working after about 20GB of writes
Hello everyone, Anybody had the chance to test out this setup and reproduce the problem? I assumed it would be something that's used often these days and a solution would benefit a lot of users. If can be of any assistance please contact me. -- Met vriendelijke groet, Richard Landsman http://rimote.nl T: +31 (0)50 - 763 04 07 (ma-vr 9:00 tot 18:00) 24/7 bij storingen: +31 (0)6 - 4388 7949 @RimoteSaS (Twitter Serviceberichten/security updates) On 04/10/2017 10:08 AM, Sandro Bonazzola wrote: Adding Paolo and Miroslav. On Sat, Apr 8, 2017 at 4:49 PM, Richard Landsman - Rimote <rich...@rimote.nl <mailto:rich...@rimote.nl>> wrote: Hello, I would really appreciate some help/guidance with this problem. First of all sorry for the long message. I would file a bug, but do not know if it is my fault, dm-cache, qemu or (probably) a combination of both. And i can imagine some of you have this setup up and running without problems (or maybe you think it works, just like i did, but it does not): PROBLEM LVM cache writeback stops working as expected after a while with a qemu-kvm VM. A 100% working setup would be the holy grail in my opinion... and the performance of KVM/qemu is great i must say in the beginning. DESCRIPTION When using software RAID 1 (2x HDD) + software RAID 1 (2xSSD) and create a cached LV out of them, the VM performs initially great (at least 40.000 IOPS on 4k rand read/write)! But then after a while (and a lot of random IO, ca 10 - 20 G) it effectively turns in to a writethrough cache although there's much space left on the cachedlv. When working as expected on KVM host all writes go to SSDs iostat -x -m 2 Device: rrqm/s wrqm/s r/s w/s rMB/swMB/s avgrq-sz avgqu-sz await r_await w_await svctm %util sda 0.00 324.500.00 22.00 0.0014.94 1390.57 1.90 86.390.00 86.39 5.32 11.70 sdb 0.00 324.500.00 22.00 0.0014.94 1390.57 2.03 92.450.00 92.45 5.48 12.05 sdc 0.00 3932.000.00 *2191.50* 0.00 *270.07* 252.3937.83 17.55 0.00 17.55 0.36 *78.05* sdd 0.00 3932.000.00 *2197.50 * 0.00 *271.01 * 252.5738.96 18.14 0.00 18.14 0.36 *78.95* When not working as expected on KVM host all writes go through the SSD on to the HDDs (effectively disabling writeback so it becomes a writethrough) Device: rrqm/s wrqm/s r/s w/s rMB/swMB/s avgrq-sz avgqu-sz await r_await w_await svctm %util sda 0.00 7.00 234.50 *173.50 * 0.92 *1.95* 14.3829.27 71.27 111.89 16.37 2.45 *100.00* sdb 0.00 3.50 212.00 *177.50 * 0.83 *1.95* 14.6035.58 91.24 143.00 29.42 2.57*100.10* sdc 2.50 0.00 566.00 *199.00 * 2.69 0.78 9.28 0.080.110.13 0.04 0.10 *7.70* sdd 1.50 0.00 76.00 *199.00* 0.65 0.78 10.66 0.020.070.16 0.04 0.07 *1.85* Stuff i've checked/tried: - The data in the cached LV has then not exceeded even half of the space, so this should not happen. It even happens when only 20% of cachedata is used. - It seems to be triggerd most of the time when %cpy/sync column of `lvs -a` is about 30%. But this is not always the case! - changing the cachepolicy from cleaner to smq, wait (check flush ready with lvs -a) and then back to smq seems to help /sometimes/! But not always... lvchange --cachepolicy cleaner /dev/mapper/XXX-cachedlv lvs -a lvchange --cachepolicy smq /dev/mapper/XXX-cachedlv - *when mounting the LV inside the host this does not seem to happen!!* So it looks like a qemu-kvm / dm-cache combination issue. Only difference is that inside host i do mkfs in stead of LVM inside VM (so could be LVM inside VM on top of LVM on KVM host problem too? small chance probably because the first 10 - 20GB it works great!) - tried disabling Selinux, upgrading to newest kernels (elrepo ml and lt), played around with dirty_cache thingeys like proc/sys/vm/dirty_writeback_centisecs /proc/sys/vm/dirty_expire_centisecs cat /proc/sys/vm/dirty_ratio , and migration threashold of dmsetup, and other probably non important stuff like vm.dirty_bytes - when in "slow state" the systems kworkers are exessively using IO (10 - 20 MB per kworker process). This seems to be the writeback process (CPY%Sync) because the cache wants to flush to HDD. But the strange thing is that after a good sync (0% left), the disk may become slow again after a few MBs of data. A reboot sometimes helps. - have tried iothreads, virtio-scsi, vcpu driver setting on virtio-scsi controller, cachesettings, disk
[CentOS-virt] lvm cache + qemu-kvm stops working after about 20GB of writes
Hello, I would really appreciate some help/guidance with this problem. First of all sorry for the long message. I would file a bug, but do not know if it is my fault, dm-cache, qemu or (probably) a combination of both. And i can imagine some of you have this setup up and running without problems (or maybe you think it works, just like i did, but it does not): PROBLEM LVM cache writeback stops working as expected after a while with a qemu-kvm VM. A 100% working setup would be the holy grail in my opinion... and the performance of KVM/qemu is great i must say in the beginning. DESCRIPTION When using software RAID 1 (2x HDD) + software RAID 1 (2xSSD) and create a cached LV out of them, the VM performs initially great (at least 40.000 IOPS on 4k rand read/write)! But then after a while (and a lot of random IO, ca 10 - 20 G) it effectively turns in to a writethrough cache although there's much space left on the cachedlv. When working as expected on KVM host all writes go to SSDs iostat -x -m 2 Device: rrqm/s wrqm/s r/s w/srMB/swMB/s avgrq-sz avgqu-sz await r_await w_await svctm %util sda 0.00 324.500.00 22.00 0.0014.94 1390.57 1.90 86.390.00 86.39 5.32 11.70 sdb 0.00 324.500.00 22.00 0.0014.94 1390.57 2.03 92.450.00 92.45 5.48 12.05 sdc 0.00 3932.000.00 *2191.50* 0.00 *270.07* 252.3937.83 17.550.00 17.55 0.36 *78.05* sdd 0.00 3932.000.00 *2197.50 *0.00 *271.01 * 252.5738.96 18.140.00 18.14 0.36 *78.95* When not working as expected on KVM host all writes go through the SSD on to the HDDs (effectively disabling writeback so it becomes a writethrough) Device: rrqm/s wrqm/s r/s w/srMB/swMB/s avgrq-sz avgqu-sz await r_await w_await svctm %util sda 0.00 7.00 234.50 *173.50 * 0.92 *1.95* 14.3829.27 71.27 111.89 16.37 2.45 *100.00* sdb 0.00 3.50 212.00 *177.50 * 0.83 *1.95* 14.6035.58 91.24 143.00 29.42 2.57*100.10* sdc 2.50 0.00 566.00 *199.00 * 2.69 0.78 9.28 0.080.110.130.04 0.10 *7.70* sdd 1.50 0.00 76.00 *199.00* 0.65 0.78 10.66 0.020.070.160.04 0.07 *1.85* Stuff i've checked/tried: - The data in the cached LV has then not exceeded even half of the space, so this should not happen. It even happens when only 20% of cachedata is used. - It seems to be triggerd most of the time when %cpy/sync column of `lvs -a` is about 30%. But this is not always the case! - changing the cachepolicy from cleaner to smq, wait (check flush ready with lvs -a) and then back to smq seems to help /sometimes/! But not always... lvchange --cachepolicy cleaner /dev/mapper/XXX-cachedlv lvs -a lvchange --cachepolicy smq /dev/mapper/XXX-cachedlv - *when mounting the LV inside the host this does not seem to happen!!* So it looks like a qemu-kvm / dm-cache combination issue. Only difference is that inside host i do mkfs in stead of LVM inside VM (so could be LVM inside VM on top of LVM on KVM host problem too? small chance probably because the first 10 - 20GB it works great!) - tried disabling Selinux, upgrading to newest kernels (elrepo ml and lt), played around with dirty_cache thingeys like proc/sys/vm/dirty_writeback_centisecs /proc/sys/vm/dirty_expire_centisecs cat /proc/sys/vm/dirty_ratio , and migration threashold of dmsetup, and other probably non important stuff like vm.dirty_bytes - when in "slow state" the systems kworkers are exessively using IO (10 - 20 MB per kworker process). This seems to be the writeback process (CPY%Sync) because the cache wants to flush to HDD. But the strange thing is that after a good sync (0% left), the disk may become slow again after a few MBs of data. A reboot sometimes helps. - have tried iothreads, virtio-scsi, vcpu driver setting on virtio-scsi controller, cachesettings, disk shedulers etc. Nothing helped. - the new samsung 950 PRO SSDs have HPA enabled (30%!!), i have AMD FX(tm)-8350, 16G RAM It feels like the lvm cache has a threshold (about 20G of data that is dirty) and that is stops allowing the qemu-kvm process to use writeback caching (the root uses inside the host seems to not have this limitation). It starts flushing, but only to a certain point. After a few MBs of data it is right back in the slow spot again. Only solution is waiting for a long time (independant of CPY%SYNC) or sometimes change cachepolicy and force flush. This prevents for me the production use of this system. But it's so promising, so I hope somebody can help. desired state: Doing the FIO test (described in section reproduce) repeatedly should keep being fast till cachedlv is more or less full. If resyncing back to disc causes this degradation, it should
Re: [CentOS-virt] Network isolation for KVM guests
Hi, I don't see why this should not work with the given solutions. But I'm relatively new to KVM / libvirt. Alternative: Personally I use Shorewall (Shoreline FW) and bridge setups (also works with a bonding interface). This way you can create zones, interfaces, addresses, forwarding-rules etc and give per VM permission to let's say only use a certain IP, only access certain parts of the network, talk to a certain limited list of IPs etc. I can not imagine you can't create what you want with Shorewall. It looks complicated, but actually is very intuitive if you give it some time and effort. Please feel free to provide a better description of what you want to accomplish. Maybe I misunderstand what you want to achieve. -- Met vriendelijke groet, Richard Landsman http://rimote.nl T: +31 (0)50 - 763 04 07 (ma-vr 9:00 tot 18:00) 24/7 bij storingen: +31 (0)6 - 4388 7949 @RimoteSaS (Twitter Serviceberichten/security updates) On 03/31/2017 11:56 AM, C. L. Martinez wrote: On Thu, Mar 30, 2017 at 06:15:28PM +0100, Nux! wrote: Use libvirt with mac/ip spoofing enabled. https://libvirt.org/formatnwfilter.html https://libvirt.org/firewall.html -- Sent from the Delta quadrant using Borg technology! Thanks Nux and Kristian but I don't see if these solutions will be really efective in my environment. Let me to explain. In this host I three physical interfaces: eth0, eth1 and wlan0. eth0 is connected to my internal network. eth1 is connected to a public router and wlan0 is connected to another public router. wlan0 and eth1 are bonded to provide failover Internet connections. CPU doesn't supports pci passthrough (pci passthrough would solve my problems). I need to deploy a fw vm to control traffic between internal and external interfaces. In BSD systems you can seggregate all ip address and route tables from principal routing table. It is the same effect that I would like to implement in this host. And I don't see how to implement using CentOS (or another linux distro). ___ CentOS-virt mailing list CentOS-virt@centos.org https://lists.centos.org/mailman/listinfo/centos-virt