Okay, this is very cool.
I updated my pseudo-production F720 for my staff to 6.4 today because I need to test out the newer code for a much bigger filer that is in production for real people :)
So, files being dumped into my ~/Maildir/ area (Maildir mail format) as mail is being delivered to me. I am the only person on staff that has this config, rest of the users use /var/mail/<user>.
I can watch my procmailrc log and see that mail is being delivered and the filenames involved.
But...
I can use 'ls' to my hearts content and I can *not* see the new messages being delivered.
ie:
uranus% ls -lt total 8580 -rw------- 1 drechsau 931 Feb 24 18:58 _BIB.L_rW-.baal -rw------- 1 drechsau 2410 Feb 24 08:29 _GcG.uwiW-.baal -rw------- 1 drechsau 12716 Feb 23 16:21 _fb.elUW-.baal -rw------- 1 drechsau 733045 Feb 23 12:35 _dj.0RRW-.baal ... uranus%
But:
From drechsau@noc.visi.com Mon Feb 24 18:59:25 2003 Subject: test Folder: /home/drechsau/Maildir/new/_pIB.t_rW-.baal 930 From procmail-bounces@lists.RWTH-Aachen.DE Mon Feb 24 18:59:37 2003 Subject: Re: Scoring problem Folder: /home/drechsau/Maildir/new/_8IB.7_rW-.baal 7811 From MAILER-DAEMON Mon Feb 24 19:00:04 2003 Subject: failure notice Folder: /home/drechsau/Maildir/new/_wKB.aAsW-.baal 5358 From MAILER-DAEMON Mon Feb 24 19:00:07 2003 Subject: Returned mail: see transcript for details Folder: /home/drechsau/Maildir/new/__KB.dAsW-.baal 8080 From mury@unknown.net Mon Feb 24 19:01:06 2003 Subject: Re: Issues coming up quick Folder: /home/drechsau/Maildir/new/_YNB.TBsW-.baal 2096 From MAILER-DAEMON Mon Feb 24 19:01:50 2003 Subject: Returned mail: see transcript for details Folder: /home/drechsau/Maildir/new/_rOB.ECsW-.baal 8683
So, files are 'there'.
Now:
uranus% ls -l /home/drechsau/Maildir/new/_rOB.ECsW-.baal -rw------- 1 drechsau 8683 Feb 24 19:01 /home/drechsau/Maildir/new/_rOB.EC sW-.baal uranus% ls -lt Maildir/new total 8580 -rw------- 1 drechsau 931 Feb 24 18:58 _BIB.L_rW-.baal -rw------- 1 drechsau 2410 Feb 24 08:29 _GcG.uwiW-.baal -rw------- 1 drechsau 12716 Feb 23 16:21 _fb.elUW-.baal -rw------- 1 drechsau 733045 Feb 23 12:35 _dj.0RRW-.baal ... uranus%
I was wondering, before I hit up netapp support, if anyone else has seen this kind of behaviour before.
I turned off nfsv3 and rebooted my workstation and I am now mounted via vers=2,proto=udp and it has been verified via nfsstat -c.
Pretty cool problem...one that will cause me to not use 6.4 as it stands right now on my full production filer that is used for mail delivery for 26K accounts in Maildir format...
On Mon, Feb 24, 2003 at 07:06:40PM -0600, Mike Horwath wrote:
I updated my pseudo-production F720 for my staff to 6.4 today because I need to test out the newer code for a much bigger filer that is in production for real people :)
So, files being dumped into my ~/Maildir/ area (Maildir mail format) as mail is being delivered to me. I am the only person on staff that has this config, rest of the users use /var/mail/<user>.
I can watch my procmailrc log and see that mail is being delivered and the filenames involved.
But...
I can use 'ls' to my hearts content and I can *not* see the new messages being delivered.
I was wondering, before I hit up netapp support, if anyone else has seen this kind of behaviour before.
I turned off nfsv3 and rebooted my workstation and I am now mounted via vers=2,proto=udp and it has been verified via nfsstat -c.
Pretty cool problem...one that will cause me to not use 6.4 as it stands right now on my full production filer that is used for mail delivery for 26K accounts in Maildir format...
Mike, I've seen this before (on Solaris AND Linux) with our F760 (a long time ago with version 5.x). ...we have Maildir's, too.
Forgive me if this seems obvious, but...
Make sure the date/time on the toaster is sync'd with the date/time on the NFS client(s)...as much as a second delay will cause this on the client(s). The difference in date/timestamps was our problem back then, and was quickly resolved (it wasn't the version of DOT that caused the problem).
--
Dave Le Blanc Unix Systems Administrator Computer Science Department California Institute of Technology
On Mon, Feb 24, 2003 at 06:06:42PM -0800, David Le Blanc wrote:
Forgive me if this seems obvious, but...
Nah, was a good idea :)
Make sure the date/time on the toaster is sync'd with the date/time on the NFS client(s)...as much as a second delay will cause this on the client(s). The difference in date/timestamps was our problem back then, and was quickly resolved (it wasn't the version of DOT that caused the problem).
Already double checked that.
Time on the filter is right on with the client and the delivery server, all NTPed together.
Tried to downgrade, oops:
RAID version 5, Ontap Version 28 expected, filesystem RAID version 5, Ontap Version 35 found
and then volume taken offline.
So I am back to 6.4 after trying to downgrade and I'll be calling into NetApp today to get this resolved.