Jianxin,Ren

@giampaolo thanks for your clarification. I am sorry you did not clearly understand what i meant as lack of more details. However, through my search, InMemoryUploadedFile object is of python...

same error with go v1.11.2 for v37.1

@zihaolucky problem above has been resolved but came with a new error `find: ‘/usr/include/python2.7’: No such file or directory` , while i use system python2.7.13 in virtualenv.

@zihaolucky above problem has been ok, but when i run sh installer.sh , came with the following error ``` ERROR: /opt/tensorboard/tensorboard-lite/tensorboard/java/org/tensorflow/tensorboard/vulcanize/BUILD:25:1: Creating runfiles tree bazel-out/host/bin/tensorboard/java/org/tensorflow/tensorboard/vulcanize/Zipper.runfiles [for host] failed (Exit 1)...

i met with similar problem about access mount as containerd Ganesha-NFS using EFS filesystem, logs resembles the following ``` Initializing Ganesha NFS server ================================== export path: /export ================================== * Writing...

@ffilz thanks, and i will make a trail but i do not well understand Ganesha NFS details especially Ganesha FSAL_VFS handle. As you suggested, is that ok whether only valid...

@ffilz hi, with DEBUG level on, more log message can be seen as follows ``` Initializing Ganesha NFS server ================================== export path: /export ================================== * Writing configuration * Starting rpcbind...

> Oops, you might want to take a look at this: > > https://patchwork.kernel.org/project/linux-fsdevel/patch/151297224519.7818.13340308932067502368.stgit@noble/ > > Apparently EFS uses a very large handle. If that can't be addressed then EFS...

In kubernetes case, both nfs-ganesha pod container and host have rpcbind service, do they conflict?

> They shouldn't, with the default network config. Each pod will have it's own network partition, so each can have it's own rpcbind. It you set the pod to use...