Skip to contentSkip to conversation

directory attribute cache on NFS4 client

DISCUSSION Ricardo Santos 3 participants 6 messages Aug 29, 2008 – Sep 3, 2008 2 messages not in the archive

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

First, some info:

NFS4 Client: Fedora 9 - 2.6.25.9-76.fc9.i686 NFS4 Server: Fedora 9 - 2.6.25.6-55.fc9.i686
exportfs: rw,fsid=0,no_root_squash,async
mount: rw,bg,retry=1,timeo=30,rsize=8192,wsize=8192

Both on a local network

The Server act as a filer to a webhosting solution. I have 2 clients (FTP and WebServer). When someone upload and overwrite a file on FTP, normally it do it:

- UPLOAD fileA.tmp
- DELETE fileA
- RENAME fileA.tmp fileA

This operation changes the inode for "fileA" and the WebServer doesn't see the changes, still getting the old file (old inode).

After upload, if I do a "ls -la /path/path/fileA", it returns the old stats (inode, mtime...). But if I do a "ls -la /path/path/" this will refresh info, and get the file correctly.

I've tried the "acdirmin" and "acdirmax" defined to 0, but didn't work. I tried too "noac", really works but the performance cost is expensive.

I really wouldn't like back to NFS3... :)

#1 lore raw

I'm checking the NFS client code to understand it, and I saw this:

-- inode.c
int nfs_attribute_timeout(struct inode *inode)

{
        struct nfs_inode *nfsi = NFS_I(inode);

        if (nfs_have_delegation(inode, FMODE_READ))
                return 0;

--

So, if a inode has a delegation, will it never be timed out, until the delegation been free, right ?

If the delegated inode has been deleted ?

A message that is not in the archive (sent off-list or not yet received): [email protected]

#3 lore raw
Quoted 10 lines · Ricardo Santos

I'm checking the NFS client code to understand it, and I saw this:

-- inode.c
int nfs_attribute_timeout(struct inode *inode)

{
        struct nfs_inode *nfsi = NFS_I(inode);

        if (nfs_have_delegation(inode, FMODE_READ))
                return 0;
Quoted 6 lines

--

So, if a inode has a delegation, will it never be timed out, until the delegation been free, right ?

If the delegated inode has been deleted ?

The server should recall the delegation before allowing the delegated file to be deleted. (The current linux server is buggy--it doesn't do that. Patches are on the way soon, I hope....)

--b.

#4 lore raw

Has anyway to disable delegations on Linux NFS Server ?

Thanks for help

A message that is not in the archive (sent off-list or not yet received): [email protected]

#6 lore raw
Quoted 1 line · Ricardo Santos

Has anyway to disable delegations on Linux NFS Server ?

yes, you can echo 0 to /proc/sys/fs/leases-enable

note that that disables all leases (atop which NFS delegations are
implemented).

	d
	.
Quoted 11 lines

Thanks for help

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#7 lore raw
Quoted 4 lines

yes, you can echo 0 to /proc/sys/fs/leases-enable

note that that disables all leases (atop which NFS delegations are implemented).

david, thanks for the tip. It really works for me!