Do you still see problems with that change?
Thanks,
Lucho
On Mar 12, 2009, at 4:07 PM, Eric Van Hensbergen wrote:
That does seem much more stable...
-eric
On Mar 12, 2009, at 11:53 AM, Abhishek Kulkarni wrote:
On Thu, Mar 12, 2009 at 11:50 AM, Eric Van Hensbergen <[email protected]
> wrote:
Anyone else see anything like this? Looks like random corruption
somewhere in the control stream between xrx and xcpufs....
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
Thu Mar 12 10:48:33 CDT 2009
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
namespace: invalid operation 9Of~
xnamespace failed: : No child processes
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
namespace: invalid operation #9PhA
xnamespace failed: : No child processes
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
namespace: invalid operation Î0A
xnamespace failed: : No child processes
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
namespace: invalid operation w§ºQ
xnamespace failed: : No child processes
[r...@arlx050 ~]# sphinx/xcpu2/utils/xrx 9.3.61.111 /bin/date
Thu Mar 12 10:48:44 CDT 2009
I can't reproduce this anymore. But does the following patch from
http://www.mail-archive.com/[email protected]/msg00068.html
fix the issue for you?
Index: xcpufs/xcpufs.c
===================================================================
--- xcpufs/xcpufs.c (revision 691)
+++ xcpufs/xcpufs.c (working copy)
@@ -1332,7 +1332,7 @@
s = p;
}
- bufwrite(&buf, buf.size, strlen(s), s);
+ bufwrite(&buf, buf.size, (slen + str - s), s);
b = 0;
bufwrite(&buf, buf.size, 1, &b);
