I applied this patch to nfs-utils-1.2.6, it fixed the current issue. My /etc/exports looks like this: /media/fs1/ 10.158.0.0/16(rw,root_squash,no_subtree_check) /media/fs1/ 10.158.192.71(rw,no_root_squash,no_subtree_check) /media/fs1/ 10.158.192.72(rw,no_root_squash,no_subtree_check) /media/fs1/ 10.158.192.73(rw,no_root_squash,no_subtree_check) /media/fs1/ 10.158.192.74(rw,no_root_squash,no_subtree_check) And I also captured dns lookup traffic by command "tcpdump -i eth0 -vvv host dns_server" during the following procedure: #exportfs -ua #exportfs -a on the server, and the nfs mount procedure on the client. There's none dns lookup related traffic, neither before the patch, nor after it. Thanks to all of you, for helping me fix this issue, and the time you spent on reviewing the code. B.R. Minlan Wang > -----Original Message----- > From: NeilBrown [mailto:neilb@xxxxxxx] > Sent: Monday, October 21, 2013 6:49 AM > To: NeilBrown > Cc: Chuck Lever; Wangminlan; J. Bruce Fields; linux-nfs@xxxxxxxxxxxxxxx > Subject: Re: Different sequence of "exportfs" produce different effects on nfs > client mounts > > > Unfortunately I cannot see how that change ever made any important > difference, and the email exchanges ends without resolving anything. > > It could only make a difference to the number of DNS lookups if there was > somewhere a test for whether clientlist[MCL_FQDN] was NULL, but there isn't > and never was. > > So it seems very likely that: > > diff --git a/support/export/client.c b/support/export/client.c index > ba2db8f..adbeed8 100644 > --- a/support/export/client.c > +++ b/support/export/client.c > @@ -767,15 +767,5 @@ client_gettype(char *ident) > sp++; > } > > - /* > - * Treat unadorned IP addresses as MCL_SUBNETWORK. > - * Everything else is MCL_FQDN. > - */ > - ai = host_pton(ident); > - if (ai != NULL) { > - freeaddrinfo(ai); > - return MCL_SUBNETWORK; > - } > - > return MCL_FQDN; > } > > is appropriate and may well fix the current issue. > > It would be good to test how many DNS looks (hopefully none) are performed > when using a exports file that contains only IP addresses, both before and after > the patch. > > NeilBrown ��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
