[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: [atomic-devel] mount gluster volume on atomic host
- From: Huamin Chen <hchen redhat com>
- To: Daniel J Walsh <dwalsh redhat com>
- Cc: atomic-devel projectatomic io
- Subject: Re: [atomic-devel] mount gluster volume on atomic host
- Date: Thu, 5 Mar 2015 17:36:07 -0500 (EST)
I am open to both.
A few things on top of my head are:
- If as an option to standard mount, we need port (and backport) Colin's patch to each mount helper (util-linux, gluster fuse xlator, cephfs, etc).
- If just use a customized SPC with this special ld.preload, we are able to avoid the above code change, though probably at the expense of some operational overhead.
----- Original Message -----
From: "Daniel J Walsh" <dwalsh redhat com>
To: "Huamin Chen" <hchen redhat com>
Cc: atomic-devel projectatomic io
Sent: Thursday, March 5, 2015 5:26:37 PM
Subject: Re: [atomic-devel] mount gluster volume on atomic host
Could we make this a general purpose tool or add this as an option to
standard mount.
On 03/05/2015 05:24 PM, Huamin Chen wrote:
> i tested the .so on nfs mount and it worked.
> ceph test is in progress - had a hard time installing it on fc21 ...
>
> ----- Original Message -----
> From: "Daniel J Walsh" <dwalsh redhat com>
> To: "Huamin Chen" <hchen redhat com>, atomic-devel projectatomic io
> Sent: Thursday, March 5, 2015 5:17:21 PM
> Subject: Re: [atomic-devel] mount gluster volume on atomic host
>
>
> On 03/05/2015 03:31 PM, Huamin Chen wrote:
>> Hi
>>
>> I applied Colin's prior util-linux patch into gluster, so atomic host can mount a gluster volume into its own namespace, through a super privileged container.
>>
>> The idea is to inject a .so into ld.preload to intercept mount(2), the .so calls setns before entering mount(2).
>>
>> The SPC is hchen/install-glusterfs-on-fc21
>>
>> Full details are here http://rootfs.github.io/Mount-Glusterfs-on-Docker-Host/
>>
>> code is here https://github.com/rootfs/install-glusterfs-on-fc21
>>
>> Huamin
>>
> If this works it is very cool. The question is how easy could it be to
> get this to work for other
> tools. (Nfs, CEPH)
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]