Search results

if you enable cache=writeback on vm, it'll enable rbd_cache=true. ceph have a feature by default rbd cache writethrough until flush = true. That mean that it's waiting to receive a first fsync, before really enable writeback. So you are safe to enable writeback. Writeback is helping for...

The problem is coming from network latency + ceph latency. If you copy 1 file, sequentially and with small blocks, it's iodepth=1. (same with dd command for example). For each block, you'll have your network latency (0,1ms for example), you'll be able to do 10000 iops. if you do it with 4k...

how do you bench it ? rados bench use multiple thread. try to test with fio, with iodepth=32 for example.

if you still have access through ssh: #apt-get install proxmox-ve should fix it

yes, it still missing support in librbd https://tracker.ceph.com/issues/20070

Hi, you can also use this external script, to backup with rbd snapshot, rbd-diff export feature. https://github.com/EnterpriseVE/eve4pve-barc Works very fine and a lot faster.

Hi, I'm using chrony too now, no more clock screw and it's also able to manage leap second. https://chrony.tuxfamily.org/comparison.html

Hi, I just see your message on https://www.frsag.org/pipermail/frsag/2018-January/009207.html, I have sent patch to fix it recently, and I should be fixed with last proxmox updates :) https://git.proxmox.com/?p=qemu-server.git;a=commit;h=87955688fda3f11440b7bc292e22409d22d8112f

Sorry, but that just mean that this specific poc only work on this "outdated" 4.9 kernel. That doesn't mean that's impossible to do the same on lasts kernels. (But yes, it's very difficult to exploit, but not impossible)

I have responded in the bugzilla, can you test to remove the socat timeout ?

ssd model ?

>>Why is "time to first byte" 2 times faster on good old 7200 rpm's in software raid-1 over Intel SSD software raid-10 you should profile your web application to see which function is doing disk access. (BTW, even 1s is quite huge for TTFB. Seem strange that it's only disk related. Are you sure...

qm is for vm , pct for containers

If the linux kernel support it, no problem. (afaik,vm memory limit is around 4TB, so it should work too with host). I never have tested more than 512G

Hi, I had time to backports my patch to last proxmox5 here a qemu-server deb build. http://odisoweb1.odiso.net/qemu-server_5.0-21_amd64.deb (need to be installed on source and destination host) then: #qm migrateexternal <vmid> [<targetip>] --targetstorage <string> (you need to copy the root...

you mean a "template" ? when a vm is transformed in template, a snasphot is done on the vm and the vm config file have a new flag "template:1" in the config file. (which made the config readonly from the gui).

here a ubuntu kernel with spectre V1, V2 retpoline && meltdown fixed (builded with gcc 7.3 to have retpoline working) https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/spectre/+build/14293994/+files/linux-image-4.13.0-33-generic_4.13.0-33.36~retpoline4_amd64.deb

Hi fabian, with spectre v2, a guest vm can read host memory. https://googleprojectzero.blogspot.be/2018/01/reading-privileged-memory-with-side.html

I'm using chrony in production, it's faster to sync clock than ntpd, openntpd. (and timesyncd is only like an cron ntpdate, really not enough precision for ceph)

similar problem with elastic and meltdown patch >>https://discuss.elastic.co/t/kernel-update-on-ubuntu-17-10-causes-elasticsearch-to-panic/116865/8