Attaching path to exportfs that:
- in client.c/client_lookup:
changes the (x)log level for unresolvable entries in
/etc/exports from L_ERROR to L_WARNING,
- in hostname.c/host_addrinfo:
changes the (x)log level for unresolvable entries in
/etc/exports from D_GENERAL to D_PARSE,
- in export.c/export_read:
counts the number of exported volume entries and
generates a (x)log L_ERROR if no volumes are exported.
Built and tested on a Fedora 20 system.
-- Henrique
Re-sending this patch, at Steve Dickson's request.
-- Henrique
--------------------------------------------------------------
Attaching path to exportfs that:
- in client.c/client_lookup:
changes the (x)log level for unresolvable entries in
/etc/exports from L_ERROR to L_WARNING,
- in hostname.c/host_addrinfo:
changes the (x)log level for unresolvable entries in
/etc/exports from D_GENERAL to D_PARSE,
- in export.c/export_read:
counts the number of exported volume entries and
generates a (x)log L_ERROR if no volumes are exported.
Built and tested on a Fedora 20 system.
--------------------------------------------------------------
Re-posting with the patch inlined with comments.
On 10/31/2014 02:33 PM, Henrique Martins wrote:
>
> Re-sending this patch, at Steve Dickson's request.
Thanks... but... :-)
There is no Signed-off-by: line as described in
https://www.kernel.org/doc/Documentation/SubmittingPatches
and there is no description section as the above document
describes
Also, please in-line the patch in the email instead
of attaching them... I just makes is easier to review.
>
> -- Henrique
>
> --------------------------------------------------------------
>
> Attaching path to exportfs that:
> - in client.c/client_lookup:
> changes the (x)log level for unresolvable entries in
> /etc/exports from L_ERROR to L_WARNING,
> - in hostname.c/host_addrinfo:
> changes the (x)log level for unresolvable entries in
> /etc/exports from D_GENERAL to D_PARSE,
> - in export.c/export_read:
> counts the number of exported volume entries and
> generates a (x)log L_ERROR if no volumes are exported.
>
> Built and tested on a Fedora 20 system.
>
> --------------------------------------------------------------
>
--- nfs-utils-1.3.0/support/export/client.c.orig 2014-09-05 08:21:37.568364360 -0700
+++ nfs-utils-1.3.0/support/export/client.c 2014-09-05 08:21:41.709451778 -0700
@@ -277,7 +277,7 @@
if (htype == MCL_FQDN && !canonical) {
ai = host_addrinfo(hname);
if (!ai) {
- xlog(L_ERROR, "Failed to resolve %s", hname);
+ xlog(L_WARNING, "Failed to resolve %s", hname);
goto out;
}
hname = ai->ai_canonname;
--- nfs-utils-1.3.0/support/export/hostname.c.orig 2014-09-05 08:09:07.387551291 -0700
+++ nfs-utils-1.3.0/support/export/hostname.c 2014-09-05 08:09:13.799573723 -0700
@@ -175,11 +175,11 @@
case 0:
return ai;
case EAI_SYSTEM:
- xlog(D_GENERAL, "%s: failed to resolve %s: (%d) %m",
+ xlog(D_PARSE, "%s: failed to resolve %s: (%d) %m",
__func__, hostname, errno);
break;
default:
- xlog(D_GENERAL, "%s: failed to resolve %s: %s",
+ xlog(D_PARSE, "%s: failed to resolve %s: %s",
__func__, hostname, gai_strerror(error));
break;
}
Just curious as to why these changes... What does D_PARSE do that
D_GENERAL does not?
--- nfs-utils-1.3.0/support/export/export.c.orig 2014-03-25 08:12:07.000000000 -0700
+++ nfs-utils-1.3.0/support/export/export.c 2014-09-05 09:23:37.424105125 -0700
@@ -76,15 +76,22 @@
struct exportent *eep;
nfs_export *exp;
+ int volumes = 0;
+
setexportent(fname, "r");
while ((eep = getexportent(0,1)) != NULL) {
exp = export_lookup(eep->e_hostname, eep->e_path, 0);
- if (!exp)
- export_create(eep, 0);
+ if (!exp) {
+ exp = export_create(eep, 0);
+ if (exp)
+ volumes++;
+ }
else
warn_duplicated_exports(exp, eep);
}
endexportent();
+ if (volumes == 0)
+ xlog(L_ERROR, "no or all unresolvable export entries");
}
/**
When testing the patch, I notice two "Failed to resolve" warning messages
are logged for each export... Its not because of this patch but
I wounder how that could be cleaned up a bit....
So please re-post with in correct patch format and we will be good to go!
steved.
Missed these comments in the middle of the patch ...
> Just curious as to why these changes... What does D_PARSE do that
> D_GENERAL does not?
D_PARSE, and L_WARNING, don't cause xlog to return an error,
D_GENERAL and L_ERROR do. An error from xlog which will
cause nfsd to error out and not start.
I changed L_ERROR with L_WARNING, but there was no
D_WARNING def or similar so I went with D_PARSE instead of
adding that one.
> When testing the patch, I notice two "Failed to resolve" warning messages
> are logged for each export... Its not because of this patch but
> I wounder how that could be cleaned up a bit....
Hopefully you solved that, otherwise I would have to hunt on
which machine I generated those patches, if I didn't wipe
out the rpmbuild environment yet..
-- Henrique