Search results

seem to be hardcoded in qemu https://git.qemu.org/?p=qemu.git;a=blob_plain;f=target/i386/cpu.c;hb=HEAD #define L2_SETS 4096 #define L3_N_SETS 16384

if you manage your own network (routers/bgp), you can change bgp annonce to failover ips to another datacenter. if you don't manage your network, for web, you can deploy reverse proxy on each datacenter and an external loadbalancer (cloudflare for example), then manage your vm ips in a private...

on my side, I have "omconfig" correctly installed /opt/dell/srvadmin/bin/omconfig (apt-get install srvadmin-all)

use the openmanage jessie repo and install the 2 missing debs manually.

yes, no problem. Just don't setup same clustername. (because multicast address is compute from clustername)

you need 2 packages from jessie to get openmanage 8.4 working on stretch currently. wget http://ftp.us.debian.org/debian/pool/main/o/openslp-dfsg/libslp1_1.2.1-10+deb8u1_amd64.deb dpkg -i libslp1_1.2.1-10+deb8u1_amd64.deb wget...

https://git.proxmox.com/?p=pve-kernel.git;a=summary 25 hours ago Fabian Grünbichler bump version to 4.13-28, bump ABI to 4.13.8-2-pve master commit | commitdiff | tree | snapshot 25 hours ago Fabian Grünbichler revert mmu changes causing bluescreens commit | commitdiff | tree | snapshot so...

It'll reduce the performance a little bit, but not too much. (virtio devices are still accelerated). I'll try to apply patches on qemu tomorrow and build a deb. if you have time to test it, it could be great.

any news for today ? just wonder if this qemu fix for hyperv and msr could help https://lists.gnu.org/archive/html/qemu-devel/2017-11/msg04221.html

yes you can change it after install. it's just add or remove some hyperv cpu flags to qemu command line.

@cybermcm : Does it crash too if you choose win2003 as windows version as os version (for you win10 guest). This disable hyperv features that qemu support. I'll like to known if it's related to hyperv or not.

the problem with dd as benchmark, is that is like iodeph=1 and sequential. so you'll be limited by the latency (network + cpu frenquency). with 18 osd ssd, replication x3 , big 2x12 cores 3.1ghz cpu, I'm able to reach around 700k iops randread 4K, and 150-200k randwrite 4K. (fio, iodepth=64...

try to add in your ceph.conf it should improve performance. They are also a bug in current luminous with debug ms, and it'll be set to 0/0 by default in next ceph release. also for rados bench, try "-t 64", to increase number of threads (16 by default)

They are a bug in bluestore currently, on memory accouting. It'll be fixed in 12.2.2 http://lists.ceph.com/pipermail/ceph-users-ceph.com/2017-October/021676.html

don't known if it's related, but new qemu patches have been sent on qemu-devel, with fix for hyperv and msr. https://lists.gnu.org/archive/html/qemu-devel/2017-11/msg04221.html

--with-local-disks on 4.0 was buggy, that's why it was fast. (you could have block corruptions). If this is slow, it's because it need to flush pending write before finish the migration. Depend of your storage speed, it can take some time. (I have done test with ssd and 10gbe network, it take...

hotplugping is on vcpus. you define cores*sockets,which is the topology and maximum vcpus you can hotplug. then you should be able to hotplug/unplug vcpus. the 80-hotplug-mem.rules need to be in the guest vm.

yes.

Hi, I'm running jewel filestore and luminous bluestore, with 3 nodes - each : 6 osd ssd - 2xintel 3ghz 12cores - 64g ram - 2x10gb ethernet . I don't have see any regression. I'm around 600000 iops randread 4k, 150000 iops randwrite 4k.

you can configure your cluster to use unicast, it should work for small cluster. (depend on latency)