My bad about the tests starting to fail at 000516, i was accidently root when I ran that test/diff.
It does in fact, like I said before fail at 000577 on down.
Brian
On Jun 29, 2004, at 9:07 PM, Brian Feeny wrote:
Sam,
Your test below does not fail, so it must be happening faster or something.
I tried to run the NFS server in sync mode, as well as disabling caching on the client using
acdirmin=0,acdirmax=0,acregmin=0,acregmax=0 in fstab. I also tried "noac" in fstab which
basically does the same thing. I also tried the NFS server with no_wdelay.
I am concerned because the NFS server is very fast (dual xeon 2.6). Do you think its ok to build
on local storage and then still use courier, or do you think its a bad sign that it didn't make it thru
the testsuite?
Also I did notice that 577 was not the first problem in the diff, here is a trail up to 577:
519d518
< 000516 * LIST (\HasNoChildren) "." "shared.test.b"
531d529
< 000519 * LIST (\HasNoChildren) "." "shared.test.b"
534d531
< 000520 * LIST (\HasNoChildren) "." "shared.test.b"
538d534
< 000521 * LIST (\HasNoChildren) "." "shared.test.b"
549d544
< 000526 * LIST (\HasNoChildren) "." "shared.test.b"
559d553
< 000529 * LIST (\HasNoChildren) "." "shared.test.b"
572d565
< 000532 * LIST (\HasNoChildren) "." "shared.test.b"
586d578
< 000535 * LIST (\HasNoChildren) "." "shared.test.b"
599d590
< 000538 * LIST (\HasNoChildren) "." "shared.test.b"
617d607
< 000543 * LIST (\HasNoChildren) "." "shared.test.b"
626d615
< 000544 * LIST (\HasNoChildren) "." "shared.test.b"
641d629
< 000548 * LIST (\Unmarked \HasNoChildren) "." "shared.test.b"
650d637
< 000549 * LIST (\Unmarked \HasNoChildren) "." "shared.test.b"
656c643
< 000554 * OK [PERMANENTFLAGS (\Draft \Answered \Flagged \Deleted \Seen)] Limited
---
> 000554 * OK [PERMANENTFLAGS ()] No permanent flags permitted
661c648
< 000559 i001 OK [READ-WRITE] Ok
---
> 000559 i001 OK [READ-ONLY] Ok
679,759c666,746
< 000577 i006 OK NOOP completed
Disabling caching didn't seem to change my output.
On Jun 29, 2004, at 8:05 PM, Sam Varshavchik wrote:
Brian Feeny writes:
Ok, just to do it, I moved my rpm build directory off NFS, and it no longer fails those tests.
Do I have to do something special to courier to get it to make it thru test on NFS?
We have ran courier for years, at least since 1.7.3 on NFS, and never had any problems.
Redhat-8.0 on the courier box, Redhat ES 2.1 on the NFS server, both running fam/portmap.
Apparently, when you move a file to a different directory (like moving a new message from the tmp directory to the new directory) if the NFS client immediately opens and read the target directory it may not see the new file right away. Looks like there's some caching going on, and the NFS client reads the cached directory contents (when the directory on the actual server contains a new file that's obviously not in the NFS client's cached copy).
Try this:
rm -rf tmp new mkdir tmp new echo "Foo" >tmp/a ls new; ln tmp/a new/a; ls new
This is not actually exactly what the part of the testsuite does, but it may be close enough to trigger the problem.
There might be some NFS tweaks to fix this. I don't know.
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
courier-users mailing list
[EMAIL PROTECTED]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
courier-users mailing list
[EMAIL PROTECTED]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
