toasters April 2000

toasters@lists.teaparty.net

121 participants
139 discussions

Possible CIFS bug?
by Marc A Levy 19 Apr '00

19 Apr '00

After the daylight savings time change a few weeks back I noticed something odd on the drives I have mapped to a NetApp. On a weekly basis (or so) I like to synchronize my home directory on the NetApp with a copy I keep on my home computer. I have been using a shareware program called "FileSync" which compares two directory trees based on file name, size, and date. There is also a option to do a binary compare (takes much longer, obviously). So, on to the problem: After the time change from EST to EDT, all of the files on the NetApp are exactly one hour older than on my home computer. Based on a little experimentation I did on the PC, I found that when the option to adjust for daylight savings time is disable on the Windows NT machine, all time appearances on the files (including the ones on filer) change by 1 hour. A article on the MSKB states that NT stores all file times in GMT, and converts the display based on the time zone setting on the PC. I am told that the NetApp stores this information in a UNIX format where it is a 32-bit integer which is "the number of seconds since Jan 1, 1970" I would suspect that CIFS converts this to GMT for compliance with Windows based on the time zone setting. It would seem there is a error in this conversion where the GMT on the file changes with the Time Zone setting (EST vs EDT). The calculation for GMT should be such that no matter what your time zone setting (even if you move your filer from EST to PST) the stamp on the file should remain constant. So, is this a bug? or some specific problem I am having? Thanks, Marc

1 0

RE:
by Bond, Andrew 19 Apr '00

19 Apr '00

You should talk to NetApp support if you are having a problem with this. However, there is nothing to do on the filer. If the autoloader is correctly configured, then using the urst0a device name will cause autoloading to occur. See the manpage for na_tape.4 for more detail. Regards, Andrew > -----Original Message----- > From: owner-dl-toasters(a)netapp.com > [mailto:owner-dl-toasters@netapp.com]On Behalf Of > bhaskar.g(a)philips.com > Sent: 19 April 2000 05:31 > To: toasters(a)mathworks.com > Subject: > > > Hi All, > > We have a Filer 720 with two shelves. We are having a DLT > 7000 Autoloader ( Dell Powervault 120T ) for the > backup of the filer. Please let me know how to enable the > autoloading option 'ON' on the filer. > > regards > Bhaskar G > Systems Engineer, > Philips Semiconductors, > #3, Sterling Square,Madras Bank Road, > Off St Mark's Road,BANGALORE-560 001. > Voice : +91 80 222 9200 Ext 1478 > Fax: +91 80 2229700 > >

1 0

None
by bhaskar.g＠philips.com 19 Apr '00

19 Apr '00

Hi All, We have a Filer 720 with two shelves. We are having a DLT 7000 Autoloader ( Dell Powervault 120T ) for the backup of the filer. Please let me know how to enable the autoloading option 'ON' on the filer. regards Bhaskar G Systems Engineer, Philips Semiconductors, #3, Sterling Square,Madras Bank Road, Off St Mark's Road,BANGALORE-560 001. Voice : +91 80 222 9200 Ext 1478 Fax: +91 80 2229700

1 0

RE: Problems with MS Excel and Oplocks?
by Day, Matt 18 Apr '00

18 Apr '00

> I've seen a strange problem over the past few days. We have MS Windows > 9x/NT users running Excel 97 and Excel 2000. Dozens of minutes, or even > hours, after a person stops editing the file (Save & Quit), you can't > access the file via NFS or CIFS. CIFS apps complain the file is in use. > NFS clients complain with "permission denied". Doing a truss of a chmod > or rm on the file shows ERR#13 - EACCES. The only way to "unlock" the > file is to have the CIFS user logoff of the client, or to terminate the > CIFS session of the Filer. The user clearly does not have any apps open > on the CIFS client. > > This sounds suspiciously like Windows 9x/NT and Excel is doing > something weird to keep the lock open. Has anyone seen this with Excel, > and how did they resolve it? That does sound like an oplock problem. What version of Data ONTAP are you running? Can you run "cifs stat" on your filer and send me the output? I'm interested in knowing what your OpLkBkNoBreakAck statistic is. You may be suffering from non-responsive oplock breaks. If you disable oplocks (set the cifs.oplocks.enable option off) and the problem goes away, then that's almost certainly the problem. I suspect a newer version of Data ONTAP will help a lot. In more recent versions there is a knob for working around the non-responsive oplock break problem (a very irritating bug in the Microsoft client). Matt

1 0

Problems with MS Excel and Oplocks?
by Timothy Demarest 18 Apr '00

18 Apr '00

I've seen a strange problem over the past few days. We have MS Windows 9x/NT users running Excel 97 and Excel 2000. Dozens of minutes, or even hours, after a person stops editing the file (Save & Quit), you can't access the file via NFS or CIFS. CIFS apps complain the file is in use. NFS clients complain with "permission denied". Doing a truss of a chmod or rm on the file shows ERR#13 - EACCES. The only way to "unlock" the file is to have the CIFS user logoff of the client, or to terminate the CIFS session of the Filer. The user clearly does not have any apps open on the CIFS client. This sounds suspiciously like Windows 9x/NT and Excel is doing something weird to keep the lock open. Has anyone seen this with Excel, and how did they resolve it? Tim -- Timothy Demarest ArrayComm, Inc. demarest(a)arraycomm.com 2480 North 1st Street, Suite 200 http://www.arraycomm.com San Jose, CA 95131

1 0

Large vs Small Backups
by Moshe Linzer 18 Apr '00

18 Apr '00

Hi, We are using Budtool 4.6.1a to backup an F760 and an F630 to a Breece Hill Q2.15 jukebox with 2 DLT7000 drives. We backup each qtree as a separate request in budtool. We started getting poor performance on the smaller qtrees, and started using the hidden ndmpd logical_find command described in Bug 12424. We recently upgraded to 5.3.5P2, and I found that this option is causing ndmp connection problems between Budtool and the filer. As a result, I have backed off to 5.3.5, where this option works without a problem. My question: How best to backup a 330GB volume with many small qtrees? Should we just backup the entire volume as one request, or continue to backup each qtree individually? Are there any problems with backing up the entire volume at once? This will allow me to dump (no pun intended) the problematic ndmpd logical_find option, as well improve general dump performance. Thanks, Moshe ----------------------------------------------------------------------------- Moshe Linzer | Mastery of Unix offers real freedom. Unix Systems Manager | The price of freedom is always dear, but National Semiconductor, Israel | I'd rather pay for my freedom than live in Phone: 972-9-970-2247 | a bitmapped, pop-up-happy dungeon like NT. Fax: 972-9-970-2001 | Email: moshel(a)nsc.com | - Thomas Scoville, Unix Review -----------------------------------------------------------------------------

2 1

nfs.v3.enable
by Todd C. Merrill 18 Apr '00

18 Apr '00

I currently have this "on." If I turn this off, what will happen to V3 clients that currently have mounts to the filer? Put another way, is this option immediate, or does it only take affect on new mounts or a reboot of the filer? Until next time... The Mathworks, Inc. 508-647-7000 x7792 3 Apple Hill Drive, Natick, MA 01760-2098 508-647-7001 FAX tmerrill(a)mathworks.com http://www.mathworks.com ---

2 1

RE: NDMP backups stopped after network upgrade
by Rainchik, Aleksandr (MED, Non GE) 18 Apr '00

18 Apr '00

Yes, we've tried "btpasswd -p -u root" and "btpasswd -p -u root -h <hostname>". With no luck at all... -----Original Message----- From: Jeff Krueger [mailto:jkrueger@qualcomm.com] Sent: Monday, April 17, 2000 7:24 PM To: Rainchik, Aleksandr (MED, Non GE) Cc: toasters(a)mathworks.com Subject: Re: NDMP backups stopped after network upgrade This could be caused by the BudTool password being out of sync. The command "btpasswd -p -u root" should allow you to change this for all your NDMP aware NetApps. You could also use "btpasswd -p -u root -h <hostname>" to change it on a per-host basis if they are different. Be sure you know what you're doing before you change the password though. It would probably be a good idea to get BudTool (Legato) support involved. -- Jeff -- ------------------------------------------------------------------------ ---- Jeff Krueger E-Mail: jeff(a)qualcomm.com NetApp File Server Lead Phone: 858-651-6709 IT Engineering and Support Fax: 858-651-6627 QUALCOMM, Incorporated Web: www.qualcomm.com From "Rainchik, Aleksandr (MED, Non GE)" on Mon, 17 Apr 2000 17:04:52 CDT: > >Hello, > >We are using Budtool 4.6.1 to backup three NetApps F740 (Ontap >5.2.1P2D10). >After our network has been down completely for about 15 minutes, we >can't >do backups any more... NDMP authorization failure for user root. Look at >the >end of this message for what we have in NetApp log files. > >So, > >a) ndmpd is running on NetApp. >b) I can telnet, or rsh to NetApp as user root just fine. >c) ndmp_hostinfo is also working perfect and I can telnet to port 10000 > on NetApp. But, when I connect there, I get 5-6 garbage letters only. >If I > telnet to port 10000 to NetApp in another department, with newer >Ontap, > I get the same garbage _and_ the string "Connection successful". Does > it make any sense to NDMP gurus? :) > >d) we've tried rebooting one NetApp and that didn't help. >e) we've tried btpasswd to create a new password for user root. Didn't >help >f) NIS is fine. I can do ypwhich or ypcat on filer... > >Any ideas? Is there any way to run ndmpd in >"verbose" or "debug" mode? > >Mon Apr 17 12:00:00 CDT [statd]: 12:00pm up 11 days, 18:13 143494298 >NFS ops, 961447 CIFS ops, 22 HTTP ops >Mon Apr 17 13:00:00 CDT [statd]: 1:00pm up 11 days, 19:13 143588178 >NFS ops, 962932 CIFS ops, 22 HTTP ops >Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 >ndmpdConnectClientAuth: MD5 authorization failure by user: root. >Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:838 >message 0x746f2061 not supported. >Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:1037 >ndmp_process_messages: unexpected reply message: 0x746f2061. >Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 >ndmpdConnectClientAuth: MD5 authorization failure by user: root. >Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:838 >message 0x746f2061 not supported. >Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:1037 >ndmp_process_messages: unexpected reply message: 0x746f2061. >Mon Apr 17 13:59:32 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 >ndmpdConnectClientAuth: MD5 authorization failure by user: root. >Mon Apr 17 14:00:00 CDT [statd]: 2:00pm up 11 days, 20:13 144027002 >NFS ops, 998979 CIFS ops, 22 HTTP ops >Mon Apr 17 14:22:28 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 >ndmpdConnectClientAuth: MD5 authorization failure by user: root. >Mon Apr 17 15:00:00 CDT [statd]: 3:00pm up 11 days, 21:13 144202255 >NFS ops, 999196 CIFS ops, 22 HTTP ops >

1 0

NDMP backups stopped after network upgrade
by Rainchik, Aleksandr (MED, Non GE) 18 Apr '00

18 Apr '00

Hello, We are using Budtool 4.6.1 to backup three NetApps F740 (Ontap 5.2.1P2D10). After our network has been down completely for about 15 minutes, we can't do backups any more... NDMP authorization failure for user root. Look at the end of this message for what we have in NetApp log files. So, a) ndmpd is running on NetApp. b) I can telnet, or rsh to NetApp as user root just fine. c) ndmp_hostinfo is also working perfect and I can telnet to port 10000 on NetApp. But, when I connect there, I get 5-6 garbage letters only. If I telnet to port 10000 to NetApp in another department, with newer Ontap, I get the same garbage _and_ the string "Connection successful". Does it make any sense to NDMP gurus? :) d) we've tried rebooting one NetApp and that didn't help. e) we've tried btpasswd to create a new password for user root. Didn't help f) NIS is fine. I can do ypwhich or ypcat on filer... Any ideas? Is there any way to run ndmpd in "verbose" or "debug" mode? Mon Apr 17 12:00:00 CDT [statd]: 12:00pm up 11 days, 18:13 143494298 NFS ops, 961447 CIFS ops, 22 HTTP ops Mon Apr 17 13:00:00 CDT [statd]: 1:00pm up 11 days, 19:13 143588178 NFS ops, 962932 CIFS ops, 22 HTTP ops Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 ndmpdConnectClientAuth: MD5 authorization failure by user: root. Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:838 message 0x746f2061 not supported. Mon Apr 17 13:00:11 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:1037 ndmp_process_messages: unexpected reply message: 0x746f2061. Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 ndmpdConnectClientAuth: MD5 authorization failure by user: root. Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:838 message 0x746f2061 not supported. Mon Apr 17 13:59:21 CDT [ndmp_session_0]: ../common/ndmpd/comm.c:1037 ndmp_process_messages: unexpected reply message: 0x746f2061. Mon Apr 17 13:59:32 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 ndmpdConnectClientAuth: MD5 authorization failure by user: root. Mon Apr 17 14:00:00 CDT [statd]: 2:00pm up 11 days, 20:13 144027002 NFS ops, 998979 CIFS ops, 22 HTTP ops Mon Apr 17 14:22:28 CDT [ndmp_session_0]: ../common/ndmpd/connect.c:196 ndmpdConnectClientAuth: MD5 authorization failure by user: root. Mon Apr 17 15:00:00 CDT [statd]: 3:00pm up 11 days, 21:13 144202255 NFS ops, 999196 CIFS ops, 22 HTTP ops

2 1

RE: How do I move /vol/vol0?
by Fox, Adam 17 Apr '00

17 Apr '00

> > I'm in a situatio where we have one F760 with 4 shelves of > 9Gb SCSI disk, and > three shelves of 36Gb FC attached disks. The FC disks were > added later, so the > original install of vol0 was done to the SCSI disks. > > We'd like to get rid of the SCSI disks (and use one shelf of > FC to hold it), > but I'm concerned since we have /vol/vol0 on it. Has anyone > moved /vol/vol0 off > SCSI to FC before, and if so, what problems did you run into? > Many customers have used vol copy to move the data over. Just make sure that the destination fiber-based volume is at least as large as the original SCSI volume. You vol copy the data over. Set the new volume to be the root volume. Then you can destroy the SCSI volume. I'd probably type download, just to be safe. Then I'd reboot...the filer won't actually destroy the old vol0 and start using the new volume until you actually reboot. Then if you want to keep the vol0 name, you can rename the new root volume. This has worked for many customers that I've talked to. Has anyone seen differently? -- Adam Fox NetApp Professional Services, NC adamfox(a)netapp.com

3 3

← Newer
1
2
3
4
5
6
7
8
...
14
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters April 2000