To stop further mailing, visit here or write:
D|RECTBUY, P.O. Box 1572, B0ULDER, CO 80306
D|RECTBUY, P.O. Box 1572, B0ULDER, CO 80306
Please see attachements. Cheers, Mickey. From 24d61cba45998815c8c2aa6fdcb8abb7fc2d85b6 Mon Sep 17 00:00:00 2001 From: Michael 'Mickey' Lauer < mic ... @vanille-mediade > Date: Tue, 10 Mar 2009 11:41:04 +0100 Subject: [PATCH] posix.vapi: rename stat struct and add actual stat + fstat calls Signed-off-by: Michael 'Mickey' Lauer < mla ... @vanille-mediade > --- vapi/posix.vapi | 6 +++++- 1 files changed, 5 insertions(+), 1 deletions(-) diff --git a/vapi/posix.vapi b/vapi/posix.vapi index 38ab008..77d7d7d 100644 --- a/vapi/posix.vapi +++ b/vapi/posix.vapi @@ -649,7 +649,7 @@ namespace Posix { public bool S_ISSOCK (mode_t mode); [CCode (cheader_filename = "sys/stat.h", cname = "struct stat")] - public struct stat { + public struct Stat { public dev_t st_dev; public ino_t st_ino; public mode_t st_mode; @@ -664,6 +664,10 @@ namespace Posix { public blksize_t st_blksize; public blkcnt_t st_blocks; } + [CCode (cheader_filename = "sys/stat.h")] + int fstat( int fd, out Stat buf); + [CCode (cheader_filename = "sys/stat.h")] + int stat (string filename, out Stat buf); [SimpleType] [IntegerType (rank = 9)] -- 1.5.6.3 From 5d2549bec07583467a4f67951a84c4c69f21e7dc Mon Sep 17 00:00:00 2001 From: Michael 'Mickey' Lauer < mic ... @vanille-mediade > Date: Tue, 10 Mar 2009 16:53:01 +0100 Subject: [PATCH] posix.vapi: signal is allowed to be null (restoring the original handler) Signed-off-by: Michael 'Mickey' Lauer < mla ... @vanille-mediade > --- vapi/posix.vapi | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/vapi/posix.vapi b/vapi/posix.vapi index 77d7d7d..203cdf8 100644 --- a/vapi/posix.vapi +++ b/vapi/posix.vapi @@ -366,7 +366,7 @@ namespace Posix { public static delegate void sighandler_t (int signal); [CCode (cheader_filename = "signal.h")] - public sighandler_t signal (int signum, sighandler_t handler); + public sighandler_t signal (int signum, sighandler_t? handler); [CCode (cheader_filename = "string.h")] public int memcmp (void* s1, void* s2, size_t n); -- 1.5.6.3 From dc9ce43ae9f42be9690de0f8d7abe2f3d6c2c3fe Mon Sep 17 00:00:00 2001 From: Michael 'Mickey' Lauer < mic ... @vanille-mediade > Date: Wed, 11 Mar 2009 11:32:54 +0100 Subject: [PATCH] glib-2.0.vapi: Add constructor for class Dir. g_dir_open() is the only valid way to obtain a new Dir structure. Without this patch, the following (legal to valac) program would compile with valac, but not with gcc: var dir = Dir(); Signed-off-by: Michael 'Mickey' Lauer < mla ... @vanille-mediade > --- vapi/glib-2.0.vapi | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/vapi/glib-2.0.vapi b/vapi/glib-2.0.vapi index 66493fa..5ca21a7 100644 --- a/vapi/glib-2.0.vapi +++ b/vapi/glib-2.0.vapi @@ -2371,7 +2371,8 @@ namespace GLib { [Compact] [CCode (free_function = "g_dir_close")] public class Dir { - public static Dir open (string filename, uint _flags = 0) throws FileError; + [CCode (cname = "g_dir_open")] + public Dir (string filename, uint _flags = 0) throws FileError; public weak string read_name (); public void rewind (); } -- 1.5.6.3 From 109912b442033cefceff17834ebf9d0857a01cac Mon Sep 17 00:00:00 2001 From: Michael 'Mickey' Lauer < mic ... @vanille-mediade > Date: Wed, 11 Mar 2009 20:17:43 +0100 Subject: [PATCH] posix.vapi: fix type for signalhandler, for posix it's called __sighandler_t, only GNU contains a typedef sighandler_t __sighandler_t. Signed-off-by: Michael 'Mickey' Lauer < mla ... @vanille-mediade > --- vapi/posix.vapi | 2 ++ 1 files changed, 2 insertions(+), 0 deletions(-) diff --git a/vapi/posix.vapi b/vapi/posix.vapi index 203cdf8..bcb59cc 100644 --- a/vapi/posix.vapi +++ b/vapi/posix.vapi @@ -363,6 +363,7 @@ namespace Posix { [CCode (cheader_filename = "signal.h")] public int kill (pid_t pid, int signum); + [CCode (cname = "__sighandler_t", cheader_filename = "signal.h")] public static delegate void sighandler_t (int signal); [CCode (cheader_filename = "signal.h")] @@ -370,6 +371,7 @@ namespace Posix { [CCode (cheader_filename = "string.h")] public int memcmp (void* s1, void* s2, size_t n); + [CCode (cheader_filename = "string.h")] public unowned string strerror (int errnum); -- 1.5.6.3 _______________________________________________ Vala-list mailing list Vala ... @ From the lack of answers I obviously have either: 1) Asked a really dumb question 2) Asked one no-one has seen as of yet or has no answer to. I will try to figure it out myself and post the results if it is #2. Also, DO NOT use NFS version 4 yet with any high-throughput production system. I am not sure of all the reasons, but Delegation doesn't work very good, nor does it failover in any decent timeframe. From our testing so far, to mount to a NAS cluster on Solaris 10 on the client use: mount -o bg,hard,timeo=50,vers=3 : James On Jun 4, 2008, at 1:51 PM, James Wonder wrote: Hello, I have setup our new NAS cluster with CFS Version 5 on Solaris systems. During testing I have found an interesting item that I need help on. In the BARG it says "The NFSRestart, NFS, and Share agents must be in same service group." which is what I did. I didn't thin k it was right to have two NFS resources, but I did it anyway. We have two systems, and two service groups which share the disks out, we do this so we can spread the load to the two servers. So, setup is as follows: 1) Solaris with Version 5 CFS 2) Two service groups, both with NFS, NIC, Share, IP and NFSRestart resources. Problem: When I fail the group to the same machine or from the same machine it stops the NFS on that machine, hence affecting the share on the other group that I am not failing or stopping. Hence, all NFS traffic is down to all my groups which is bad. I tried making a proxy to an NFS resource in a Parallel group, the same problem happens. I was going to make the Proxy agent "MonitorOnly", but thought better before I emailed. This makes sense, and what I thought would happen, but I still am worried about the BARG comment. Any help is appreciated!!!! James _______________________________________________ Veritas-ha maillist - Veri ... @ /veritas-ha _______________________________________________ Veritas-ha maillist - Veri ... @
