Bug 20774
| Summary: | 'NFS access cache time=2' is not a daemon | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Base System | Reporter: | Andre Albsmeier <Andre.Albsmeier> | ||||
| Component: | conf | Assignee: | Doug Barton <dougb> | ||||
| Status: | Closed FIXED | ||||||
| Severity: | Affects Only Me | ||||||
| Priority: | Normal | ||||||
| Version: | 4.1-STABLE | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| Attachments: |
|
||||||
Responsible Changed From-To: freebsd-bugs->sheldonh This one was sent for my attention. On Tue, 22 Aug 2000 08:54:13 +0200, Andre Albsmeier wrote:
> Since 'NFS access cache time=2' is not a daemon we might shut it up
> or move it somewhere else, e.g.:
>
> Additional network settings: NFS access cache time=2
Do you agree, though, that for the purpose of backward compatibility,
the test for whether or not to frob the sysctl should test not only the
value of nfs_access_cache, but also the value of nfs_client_enable?
i.e.
echo -n 'Additional network settings:'
case ${nfs_client_enable} in
[Yy][Ee][Ss])
if [ -n "${nfs_access_cache}" ]; then
echo -n " NFS access cache time=${nfs_access_cache}"
sysctl -w vfs.nfs.access_cache_timeout=${nfs_access_cache} > /dev/null
fi
;;
esac
echo '.'
Ciao,
Sheldon.
On Tue, 22-Aug-2000 at 16:43:25 +0200, Sheldon Hearn wrote: > > > On Tue, 22 Aug 2000 08:54:13 +0200, Andre Albsmeier wrote: > > > Since 'NFS access cache time=2' is not a daemon we might shut it up > > or move it somewhere else, e.g.: > > > > Additional network settings: NFS access cache time=2 > > Do you agree, though, that for the purpose of backward compatibility, > the test for whether or not to frob the sysctl should test not only the > value of nfs_access_cache, but also the value of nfs_client_enable? Yes. It would work even without testing nfs_client_enable but would probably fail if no NFS support was compiled into the kernel -Andre > > i.e. > > echo -n 'Additional network settings:' > > case ${nfs_client_enable} in > [Yy][Ee][Ss]) > > if [ -n "${nfs_access_cache}" ]; then > echo -n " NFS access cache time=${nfs_access_cache}" > sysctl -w vfs.nfs.access_cache_timeout=${nfs_access_cache} > /dev/null > fi > ;; > esac > > echo '.' > > Ciao, > Sheldon. On Tue, 22 Aug 2000 17:38:24 +0200, Andre Albsmeier wrote:
> Yes. It would work even without testing nfs_client_enable but
> would probably fail if no NFS support was compiled into the kernel
I had a look and decided that it's not really important to limit the
output associated with "starting daemons" to daemon names. It's
perfectly acceptable to include output relating to the operation of
those daemons. I think. :-)
However, Bruce Evans' rev 1.46 of src/etc/rc lured me in, and I think
that something like this might be helpful.
Bruce, is this what you had in mind?
Ciao,
Sheldon.
Index: rc.network
===================================================================
RCS file: /home/ncvs/src/etc/rc.network,v
retrieving revision 1.83
diff -u -d -r1.83 rc.network
--- rc.network 2000/08/16 23:08:28 1.83
+++ rc.network 2000/08/22 16:19:20
@@ -8,6 +8,11 @@
# first before contemplating any changes here. If you do need to change
# this file for some reason, we would like to know about it.
+
+v_set_sysctl() {
+ echo -n " '`sysctl -w $1 2>&1`'"
+}
+
# First pass startup stuff.
#
network_pass1() {
@@ -270,85 +275,73 @@
[Yy][Ee][Ss] | '')
;;
*)
- echo -n ' tcp extensions=NO'
- sysctl -w net.inet.tcp.rfc1323=0 >/dev/null
+ v_set_sysctl net.inet.tcp.rfc1323=0
;;
esac
case ${icmp_bmcastecho} in
[Yy][Ee][Ss])
- echo -n ' broadcast ping responses=YES'
- sysctl -w net.inet.icmp.bmcastecho=1 >/dev/null
+ v_set_sysctl net.inet.icmp.bmcastecho=1
;;
esac
case ${icmp_drop_redirect} in
[Yy][Ee][Ss])
- echo -n ' ignore ICMP redirect=YES'
- sysctl -w net.inet.icmp.drop_redirect=1 >/dev/null
+ v_set_sysctl net.inet.icmp.drop_redirect=1
;;
esac
case ${icmp_log_redirect} in
[Yy][Ee][Ss])
- echo -n ' log ICMP redirect=YES'
- sysctl -w net.inet.icmp.log_redirect=1 >/dev/null
+ v_set_sysctl net.inet.icmp.log_redirect=1
;;
esac
case ${gateway_enable} in
[Yy][Ee][Ss])
- echo -n ' IP gateway=YES'
- sysctl -w net.inet.ip.forwarding=1 >/dev/null
+ v_set_sysctl net.inet.ip.forwarding=1
;;
esac
case ${forward_sourceroute} in
[Yy][Ee][Ss])
- echo -n ' do source routing=YES'
- sysctl -w net.inet.ip.sourceroute=1 >/dev/null
+ v_set_sysctl net.inet.ip.sourceroute=1
;;
esac
case ${accept_sourceroute} in
[Yy][Ee][Ss])
- echo -n ' accept source routing=YES'
- sysctl -w net.inet.ip.accept_sourceroute=1 >/dev/null
+ v_set_sysctl net.inet.ip.accept_sourceroute=1
;;
esac
case ${tcp_keepalive} in
[Yy][Ee][Ss])
- echo -n ' TCP keepalive=YES'
- sysctl -w net.inet.tcp.always_keepalive=1 >/dev/null
+ v_set_sysctl net.inet.tcp.always_keepalive=1
;;
esac
case ${tcp_restrict_rst} in
[Yy][Ee][Ss])
- echo -n ' restrict TCP reset=YES'
- sysctl -w net.inet.tcp.restrict_rst=1 >/dev/null
+ v_set_sysctl net.inet.tcp.restrict_rst=1
;;
esac
case ${tcp_drop_synfin} in
[Yy][Ee][Ss])
- echo -n ' drop SYN+FIN packets=YES'
- sysctl -w net.inet.tcp.drop_synfin=1 >/dev/null
+ v_set_sysctl net.inet.tcp.drop_synfin=1
;;
esac
case ${ipxgateway_enable} in
[Yy][Ee][Ss])
- echo -n ' IPX gateway=YES'
- sysctl -w net.ipx.ipx.ipxforwarding=1 >/dev/null
+ v_set_sysctl net.ipx.ipx.ipxforwarding=1
;;
esac
case ${arpproxy_all} in
[Yy][Ee][Ss])
- echo -n ' ARP proxyall=YES'
- sysctl -w net.link.ether.inet.proxyall=1 >/dev/null
+ v_set_sysctl net.link.ether.inet.proxyall=1
;;
esac
@@ -356,8 +349,7 @@
[Nn][Oo] | '')
;;
*)
- echo -n ' ip_portrange_first=$ip_portrange_first'
- sysctl -w net.inet.ip.portrange.first=$ip_portrange_first >/dev/null
+ v_set_sysctl net.inet.ip.portrange.first=$ip_portrange_first
;;
esac
@@ -365,8 +357,7 @@
[Nn][Oo] | '')
;;
*)
- echo -n ' ip_portrange_last=$ip_portrange_last'
- sysctl -w net.inet.ip.portrange.last=$ip_portrange_last >/dev/null
+ v_set_sysctl net.inet.ip.portrange.last=$ip_portrange_last
;;
esac
@@ -527,15 +518,14 @@
case ${nfs_reserved_port_only} in
[Yy][Ee][Ss])
- echo -n ' NFS on reserved port only=YES'
- sysctl -w vfs.nfs.nfs_privport=1 >/dev/null
+ v_set_sysctl vfs.nfs.nfs_privport=1
;;
esac
echo -n ' nfsd'; nfsd ${nfs_server_flags}
if [ -n "${nfs_bufpackets}" ]; then
- sysctl -w vfs.nfs.bufpackets=${nfs_bufpackets} \
+ v_set_sysctl vfs.nfs.bufpackets=${nfs_bufpackets} \
> /dev/null
fi
@@ -575,9 +565,7 @@
[Yy][Ee][Ss])
echo -n ' nfsiod'; nfsiod ${nfs_client_flags}
if [ -n "${nfs_access_cache}" ]; then
- echo -n " NFS access cache time=${nfs_access_cache}"
- sysctl -w vfs.nfs.access_cache_timeout=${nfs_access_cache} \
- >/dev/null
+ v_set_sysctl vfs.nfs.access_cache_timeout=${nfs_access_cache}
fi
;;
esac
@@ -672,9 +660,8 @@
[Nn][Oo] | '')
;;
*)
- echo -n ' log_in_vain=YES'
- sysctl -w net.inet.tcp.log_in_vain=1 >/dev/null
- sysctl -w net.inet.udp.log_in_vain=1 >/dev/null
+ v_set_sysctl net.inet.tcp.log_in_vain=1
+ v_set_sysctl net.inet.udp.log_in_vain=1
;;
esac
On Tue, 22-Aug-2000 at 18:23:49 +0200, Sheldon Hearn wrote: > > > On Tue, 22 Aug 2000 17:38:24 +0200, Andre Albsmeier wrote: > > > Yes. It would work even without testing nfs_client_enable but > > would probably fail if no NFS support was compiled into the kernel > > I had a look and decided that it's not really important to limit the > output associated with "starting daemons" to daemon names. It's > perfectly acceptable to include output relating to the operation of > those daemons. I think. :-) Well, I always liked the idea of having one word of diagnostics for each deamon which is easy to parse. So, if something goes wrong, the error messages directly jump into ones eye (I hope this can be expressed like this in english). I always found the NFS cache timeout thing a bit confusing regarding to this but this is probably my personal opinion only. I even don't know why this thing has to be set. The default value (set in the system) is 60 but as soon as the machine is an NFS client it gets set to 2. Without being an NFS client it probably doesn't matter so if 2 is the appropriate value it could be set to this by default. But this is a different topic... -Andre > > However, Bruce Evans' rev 1.46 of src/etc/rc lured me in, and I think > that something like this might be helpful. > > Bruce, is this what you had in mind? > > Ciao, > Sheldon. > > Index: rc.network > =================================================================== > RCS file: /home/ncvs/src/etc/rc.network,v > retrieving revision 1.83 > diff -u -d -r1.83 rc.network > --- rc.network 2000/08/16 23:08:28 1.83 > +++ rc.network 2000/08/22 16:19:20 > @@ -8,6 +8,11 @@ > # first before contemplating any changes here. If you do need to change > # this file for some reason, we would like to know about it. > > + > +v_set_sysctl() { > + echo -n " '`sysctl -w $1 2>&1`'" > +} > + > # First pass startup stuff. > # > network_pass1() { > @@ -270,85 +275,73 @@ > [Yy][Ee][Ss] | '') > ;; > *) > - echo -n ' tcp extensions=NO' > - sysctl -w net.inet.tcp.rfc1323=0 >/dev/null > + v_set_sysctl net.inet.tcp.rfc1323=0 > ;; > esac > > case ${icmp_bmcastecho} in > [Yy][Ee][Ss]) > - echo -n ' broadcast ping responses=YES' > - sysctl -w net.inet.icmp.bmcastecho=1 >/dev/null > + v_set_sysctl net.inet.icmp.bmcastecho=1 > ;; > esac > > case ${icmp_drop_redirect} in > [Yy][Ee][Ss]) > - echo -n ' ignore ICMP redirect=YES' > - sysctl -w net.inet.icmp.drop_redirect=1 >/dev/null > + v_set_sysctl net.inet.icmp.drop_redirect=1 > ;; > esac > > case ${icmp_log_redirect} in > [Yy][Ee][Ss]) > - echo -n ' log ICMP redirect=YES' > - sysctl -w net.inet.icmp.log_redirect=1 >/dev/null > + v_set_sysctl net.inet.icmp.log_redirect=1 > ;; > esac > > case ${gateway_enable} in > [Yy][Ee][Ss]) > - echo -n ' IP gateway=YES' > - sysctl -w net.inet.ip.forwarding=1 >/dev/null > + v_set_sysctl net.inet.ip.forwarding=1 > ;; > esac > > case ${forward_sourceroute} in > [Yy][Ee][Ss]) > - echo -n ' do source routing=YES' > - sysctl -w net.inet.ip.sourceroute=1 >/dev/null > + v_set_sysctl net.inet.ip.sourceroute=1 > ;; > esac > > case ${accept_sourceroute} in > [Yy][Ee][Ss]) > - echo -n ' accept source routing=YES' > - sysctl -w net.inet.ip.accept_sourceroute=1 >/dev/null > + v_set_sysctl net.inet.ip.accept_sourceroute=1 > ;; > esac > > case ${tcp_keepalive} in > [Yy][Ee][Ss]) > - echo -n ' TCP keepalive=YES' > - sysctl -w net.inet.tcp.always_keepalive=1 >/dev/null > + v_set_sysctl net.inet.tcp.always_keepalive=1 > ;; > esac > > case ${tcp_restrict_rst} in > [Yy][Ee][Ss]) > - echo -n ' restrict TCP reset=YES' > - sysctl -w net.inet.tcp.restrict_rst=1 >/dev/null > + v_set_sysctl net.inet.tcp.restrict_rst=1 > ;; > esac > > case ${tcp_drop_synfin} in > [Yy][Ee][Ss]) > - echo -n ' drop SYN+FIN packets=YES' > - sysctl -w net.inet.tcp.drop_synfin=1 >/dev/null > + v_set_sysctl net.inet.tcp.drop_synfin=1 > ;; > esac > > case ${ipxgateway_enable} in > [Yy][Ee][Ss]) > - echo -n ' IPX gateway=YES' > - sysctl -w net.ipx.ipx.ipxforwarding=1 >/dev/null > + v_set_sysctl net.ipx.ipx.ipxforwarding=1 > ;; > esac > > case ${arpproxy_all} in > [Yy][Ee][Ss]) > - echo -n ' ARP proxyall=YES' > - sysctl -w net.link.ether.inet.proxyall=1 >/dev/null > + v_set_sysctl net.link.ether.inet.proxyall=1 > ;; > esac > > @@ -356,8 +349,7 @@ > [Nn][Oo] | '') > ;; > *) > - echo -n ' ip_portrange_first=$ip_portrange_first' > - sysctl -w net.inet.ip.portrange.first=$ip_portrange_first >/dev/null > + v_set_sysctl net.inet.ip.portrange.first=$ip_portrange_first > ;; > esac > > @@ -365,8 +357,7 @@ > [Nn][Oo] | '') > ;; > *) > - echo -n ' ip_portrange_last=$ip_portrange_last' > - sysctl -w net.inet.ip.portrange.last=$ip_portrange_last >/dev/null > + v_set_sysctl net.inet.ip.portrange.last=$ip_portrange_last > ;; > esac > > @@ -527,15 +518,14 @@ > > case ${nfs_reserved_port_only} in > [Yy][Ee][Ss]) > - echo -n ' NFS on reserved port only=YES' > - sysctl -w vfs.nfs.nfs_privport=1 >/dev/null > + v_set_sysctl vfs.nfs.nfs_privport=1 > ;; > esac > > echo -n ' nfsd'; nfsd ${nfs_server_flags} > > if [ -n "${nfs_bufpackets}" ]; then > - sysctl -w vfs.nfs.bufpackets=${nfs_bufpackets} \ > + v_set_sysctl vfs.nfs.bufpackets=${nfs_bufpackets} \ > > /dev/null > fi > > @@ -575,9 +565,7 @@ > [Yy][Ee][Ss]) > echo -n ' nfsiod'; nfsiod ${nfs_client_flags} > if [ -n "${nfs_access_cache}" ]; then > - echo -n " NFS access cache time=${nfs_access_cache}" > - sysctl -w vfs.nfs.access_cache_timeout=${nfs_access_cache} \ > - >/dev/null > + v_set_sysctl vfs.nfs.access_cache_timeout=${nfs_access_cache} > fi > ;; > esac > @@ -672,9 +660,8 @@ > [Nn][Oo] | '') > ;; > *) > - echo -n ' log_in_vain=YES' > - sysctl -w net.inet.tcp.log_in_vain=1 >/dev/null > - sysctl -w net.inet.udp.log_in_vain=1 >/dev/null > + v_set_sysctl net.inet.tcp.log_in_vain=1 > + v_set_sysctl net.inet.udp.log_in_vain=1 > ;; > esac > -- Windows NT Multitasking: Messing up several things at once. On Tue, 22 Aug 2000, Sheldon Hearn wrote: > On Tue, 22 Aug 2000 17:38:24 +0200, Andre Albsmeier wrote: > > > Yes. It would work even without testing nfs_client_enable but > > would probably fail if no NFS support was compiled into the kernel > > I had a look and decided that it's not really important to limit the > output associated with "starting daemons" to daemon names. It's > perfectly acceptable to include output relating to the operation of > those daemons. I think. :-) In old mail to Andre about this, I said that the daemon-related output should be together with the daemon names. Now I think s separate section is better. There is too much output to format nicely. > However, Bruce Evans' rev 1.46 of src/etc/rc lured me in, and I think rc.network > that something like this might be helpful. > > Bruce, is this what you had in mind? Not quite. I didn't think of your good idea of using a function. Maybe the function could format the output nicely (somehow mark up the output associated with daemons, and wrap lines nicely...). Bruce What about doing something like this?
1) add a helper function
v_echo_status() {
echo -n " ['$1']";
}
2) adjust rc output
- echo -n " NFS access cache time=${nfs_access_cache}"
+ v_echo_status "NFS access cache time=${nfs_access_cache}"
which will produce:
nfsiod [NFS access cache time=2]
This intermingles daemon names and their associated status/options, but is
in a format that is easily parseable by humans and machines.
--
Matt Emmerton
State Changed From-To: open->suspended We'll rethink this one once the rcNG process has settled. Right now, it's barely started. Responsible Changed From-To: sheldonh->dougb Ugh, forgot to hand this over to dougb, who's overseeing the rcNG process and could be considered the startup scripts' maintainer. State Changed From-To: suspended->closed While I agree that this isn't as clean as it could be, the nfs access cache time is related to the nfs client stuff, so it's not totally out of place here. The alternative solutions make the code more complicated, for no great benefit. In the long run, with any kind of luck, all this code is going away, so this level of change for little benefit doesn't seem justified. |
When booting a NFS client, the following line is spit out: Starting final network daemons: mountd nfsd nfsiod NFS access cache time=2 amd rwhod. It is generated from these lines in /etc/rc.network: case ${nfs_client_enable} in [Yy][Ee][Ss]) echo -n ' nfsiod'; nfsiod ${nfs_client_flags} if [ -n "${nfs_access_cache}" ]; then echo -n " NFS access cache time=${nfs_access_cache}" sysctl -w vfs.nfs.access_cache_timeout=${nfs_access_cache} \ >/dev/null fi ;; esac Since 'NFS access cache time=2' is not a daemon we might shut it up or move it somewhere else, e.g.: Additional network settings: NFS access cache time=2 Fix: Patch for moving it to an additional section: How-To-Repeat: Boot a NFS client.