Jian Wang

Results 533 comments of Jian Wang

In http://kubevirt.io/api-reference/v0.55.0/definitions.html (as of 2022.07) , the `ide` is not in `bus type`. ![image](https://user-images.githubusercontent.com/31133476/180857571-10079ba0-099a-453d-930e-4dda053fbaca.png)

More clues: The root `/dev/mapper/centos-root` is not successfully mapped via `dev-mapper` when running via kubevirt. Not sure if it is not mapped, or mapped failed, or other reasons. cat `/proc/cmdline`...

connect with https://github.com/harvester/harvester/issues/2236 SRIOV may also be represented as a kind of network to reuse existing features

@DaiYuzeng I would like to suggest that we add similar resource configuration for Grafana POD in Harvester GUI. This issue is in lower priority than others.

@WuJun2016 @joshmeranda @guangbochen

Harvester has Longhorn as the CSI layer, by default it will have 3 replicas for PVC, in your setup, there are 3 nodes, each node will hold one replica. When...

@fanlix It could be such a case: Your application in guest VM, uses roughly 64 GB memory in peak (OS + APP) , and then, the APP releases memory, thus...

roughly, except the `seems free` part, all other kinds of memory in guest VM, are counted as the really used memory in HOST. OOM happend to HOST side. ![image](https://user-images.githubusercontent.com/31133476/176855043-5acebb30-d8ca-4ef0-989e-749dc7e8e338.png)

Not sure how your APP really manage the memory. ``` this program using jemalloc, and alloc/free huge memory very fast. ``` `memory hole` is the suspecting point. Physical machine can...

> https://github.com/harvester/harvester/issues/2419#issuecomment-1207977489 try 8G ,4G, 2G, 1G, 512M, 256M, all pass. 128M,100M cause oom. Thanks @fanlix , the experiment is fairly valuable. Connect with #2522 , the reserved memory, deserves...