Re: [pve-devel] transparent huge pages support / disk passthrough corruption

It seems that the current implementation is much better than it was in the RHEL-based kernel. On Thu, Jan 19, 2017 at 9:43 AM, Alexandre DERUMIER wrote: > Hi, > > I have reenable THP ( transparent_hugepage=madvise) since around 1 year > (with pve-kernel 4.2-4.4), and I

Re: [pve-devel] transparent huge pages support / disk passthrough corruption

Hi, I have reenable THP ( transparent_hugepage=madvise) since around 1 year (with pve-kernel 4.2-4.4), and I don't have problem anymore like in the past. I'm hosting a lot of database (mysql,sqlserver, redis, mongo,...) and I don't have seen performance impact since I have reenable THP. So I

[pve-devel] transparent huge pages support / disk passthrough corruption

So it seems like the recently reported problems[1] with disk pass through using virtio-scsi(-single) are caused by a combination of Qemu since 2.7 not handling memory fragmentation (well) and our compiled-in default of disabling transparent huge pages on the kernel side. While I will investigate