Hello Sir,<br />
<br />
I'm working in an large Linux environment. We have lots of CentOS servers (different versions).<br />
<br />
We used to partition one large virtual disk derived from disk array to several 1TB partitions, to avoid tremendous fsck time and quotacheck time.<br />
<br />
In this case we'll got many shared filesystems from one server, which is hard to remember, so we create one mountpoint under /, and create several sub- mountpoints to mount these filesystems, like this:<br />
<br />
/total<br />
/total/total_1<br />
/total/total_2<br />
/total/total_3<br />
<br />
This is where the NFS options "crossmnt" and "nohide" do help.<br />
<br />
Here is one of my /etc/exports:<br />
<br />
/hv-test-auto-run *(rw,async,insecure,insecure_locks,crossmnt,fsid=0)<br />
/hv-test-auto-run/hv-test-auto-run1 *(rw,async,insecure,insecure_locks,nohide,fsid=1)<br />
/hv-test-auto-run/hv-test-auto-run2 *(rw,async,insecure,insecure_locks,nohide,fsid=2)<br />
/hv-test-auto-run/hv-test-auto-run3 *(rw,async,insecure,insecure_locks,nohide,fsid=3)<br />
/hv-test-auto-run/hv-test-auto-run4 *(rw,async,insecure,insecure_locks,nohide,fsid=4)<br />
/hv-test-auto-run/hv-test-auto-run5 *(rw,async,insecure,insecure_locks,nohide,fsid=5)<br />
<br />
<br />
Recently after I installed and configured two CentOS 5.8 hosts, I found that these 5.8 hosts can't access the sub-filesystems of another CentOS 5.4, but the 5.4 host can access the sub-filesystems of these 5.8 host, and these two 5.8 hosts can access the sub-filesystems of each other.<br />
<br />
I can guarantee that:<br />
<br />
1. There are no platform issue because I tested on servers, PCs and virtual machines.<br />
2. There are no configuration error because I asked two of my co-workers to check it with me.
↧