toasters October 1998

toasters@lists.teaparty.net

60 participants
75 discussions

raid scrubbing and hot spares
by Dave Cole 05 Nov '98

05 Nov '98

We have a number of f230's here that seem to have either stopped doing their raid scrubbing or isn't bothering to report it any more. while the /etc/rc file doesn't explicitly ask to turn on scrubbing, doing an 'rsh netapp1 options' returns this: raidtimeout 24 raid.reconstruct_speed 4 raid.scrub.enable on which seems pretty typical. Only 2 out of our 7 operating netapps are actually logging this: Sun Oct 18 01:00:00 EDT [raid_scrub_admin]: Beginning disk scrubbing... Sun Oct 18 02:47:16 EDT [raid_scrub_admin]: Scrub found 0 parity inconsistencies Sun Oct 18 02:47:16 EDT [raid_scrub_admin]: Scrub found 0 media errors Sun Oct 18 02:47:16 EDT [raid_scrub_admin]: Disk scrubbing finished... Of the two that are logging, one is an f210 running ONTAP 4.2a, the other is an f210 running ONTAP 5.1.2 The ones not logging are 4 x f230's, 1 x f220. On a related note, we happen to know that one of our hot spares on one of the f210's is showing a read error. It was discovered during the install. I'm not entirely sure if I have logs of the discovery. But along that line, are there any tests or scrubs done or are they assumed to be healthy? ---------------------------------------------------------------- Dave Cole (DC1110) | dacole(a)netcom.ca Systems Administrator |* dacole(a)rik.net * | office/~dacole/ Netcom Canada |* www.rik.net/~dacole/ * 905 King Street West, Toronto, M6K 3G9 | phone - 416.341.5801 Toronto, Ontario, Canada, Earth, Sol | fax - 416.341.5725

11 13

Changing which volume is the root volume W/O losing data?
by Larry Rosenman-CyberRamp System Administration 02 Nov '98

02 Nov '98

Is it possible on 5.1D14 to change which of 2 volumes is the root volume of a F520-110? I have a filer that has 14 disks (100GB), and 2 volumes: 30gb Mail 70gb News. Mail is the current root volume. I want to make the news volume the root volume, but not lose the data on news. I then want to destroy mail (it's moving to a new filer), and add it's disks to the news volume. How would I go about this? Thanks! LER -- Larry Rosenman, Sr. System Administrator, CyberRamp Internet Services E-Mail: ler(a)cyberramp.net, http://www.cyberramp.net Voice: (214) 343-3333/(817) 461-8484 (Metro)/Fax: (214) 343-3727 Tech Support: (214) 340-2020/(817) 226-2020 (Metro) 7AM-Midnight, 7 Days U.S. Mail: 11350 Hillguard Rd, Dallas, TX 75243-8311

3 3

Re: SMTP MIB (was Re: raid scrubbing and hot spares)
by mds＠ns.gbnet.net 29 Oct '98

29 Oct '98

Has anyone thought of adding a counter for the number of NVRAM flushes for a given rolling period? My understanding is that one of the issues in achieving excellent performance on writes is avoiding, or dealing with, back-to-back consistency points. It strikes me that a simple metric based on the number of NVRAM flushes-to-disk over the last 60-seconds, updated every 5 seconds say, would give people a handle on whether there's a bottleneck happening there. It might be even more interesting to keep track of in the clustered-filer environment.

2 1

Re: ndmpcopy question
by sirbruce＠ix.netcom.com 29 Oct '98

29 Oct '98

On 10/28/98 17:49:05 you wrote: >NOW BEWARE! Once you test a few times and discover that ndmpcopy tends to >hang, you can get into the habit of just interrupting it whenever you see >that message. DON'T. If you are copying a very large amount of data, the >restore will probably finish well after the dump. This is because the >restore does a final pass (after all data has transferred) to set the >correct owner and permissions on all directories. If you interrupt >ndmpcopy while this is going on, you are left with a bunch of dirs owned >by root, with permissions 777 on the restore filer. Can you guess how >I found this out?? :-) This should be entered as a bug. Even though interrupting a restore isn't something you should do, having it create potential security issues as a result isn't really necessary. It's probably nice to have that as a with ndmpcopy as a shortcut in the code (you can probably rely letting the requests go through a higher level, since they have "permission"), though, so it's not surprising it's there. May not be something that can be fixed easily. Bruce

1 0

Re: raid scrubbing and hot spares
by sirbruce＠ix.netcom.com 29 Oct '98

29 Oct '98

On 10/28/98 09:28:15 you wrote: > >> If I recall correctly, NetApp suggests you don't cron rdate's to >> run on the hour because it's possible to step the clock past the >> hour and miss a scheduled event such as raid scrubbing or >> snapshots. I beleive I read this on NOW, not sure. >Interesting! >Wish they would provide an alternate solution. >The don't provide NTP and then they discourage us from cron'ing rdate. >What do they think? I have nothing else meaningful to do, but sit here >and periodically do > remoteshell filer1 "hello brain dead, here is your date/time" > remoteshell filre2 "hello brain dead, here is your date/time" > .... An understandable reaction, but slightly misguided. *frequent* rdates, particularly those at the top or "common" offsets of the hour, when another cron event is to take place, are what breaks it. Doing a infrequent rdate, say 4 times an hour at odd offsets like 12, 27, 42, and 57 minutes, should be just fine. However, this is a problem that does need to be fixed once and for all. Most people probably never experience it. Bruce

1 0

ndmpcopy port for HP/UX
by Garth T Kidd 29 Oct '98

29 Oct '98

Has anyone compiled ndmpcopy for HP/UX? Regards, Garth. -- Garth T Kidd <garth(a)netapp.com> Network Appliance: Systems Engineer, Australia & NZ Fast. Simple. Reliable. Multiprotocol. Mobile: +61-414-300-213 http://www.netapp.com/

2 1

RE: raid scrubbing and hot spares
by sirbruce＠ix.netcom.com 28 Oct '98

28 Oct '98

On 10/27/98 13:48:40 you wrote: > > >Hmm... That's interesting. Is there a CPU usage threshold before the filer >becomes "Busy"? Is it 60%, 70% or what? It's probably "busy" in the sense of heavy disk activity. Which might be associated with high CPU usage but not necessarily. As someone else noted, frequent rdates can do it too. Perhaps running a backup? I know certain models and certain OSes are less prone to this, but I don't know the details of how Netapp currently views the situation. Bruce

8 8

samba and CIFS on netapp
by Bryan Hess 26 Oct '98

26 Oct '98

Recently a couple of us took a look at why Samba clients can't connect to NetApp filers running CIFS. Even something as simple as this fails: smbclient -L filername Netapp does not support samba clients, but we're interested in finding out what goes wrong. Attached to the end of this note is a summary of SMB interaction with the netapp filer, both from a windows machine and from samba. It was prepared locally by Larry Dootlittle. We'd like to see support for samba in OnTap. I'm told that it may be coming in a later version. --Bryan Hess ================================================================================ Attempt to diagnose why Samba fails to establish SMB session with a NetApp. Larry Doolittle <ldoolitt(a)jlab.org> Oct 23, 1998 Based on output of tcpdump-3.4a5 with smb patches by A. Tridell "Leach" is draft-leach-cifs-v1-spec-01.txt (expired). Windows 4 Samba Session Request (No mention of this packet in Leach) Flags 0x81000044 0x81000048 Source "WU" "RECYCLE" Session Granted (No mention of this packet in Leach) (no differences) SMBnegprot (REQUEST) Proc ID 28498 8593 UID 0 313 MID 2177 8693 SMBnegprot (REPLY) Proc ID 28498 8593 UID 0 0 MID 2177 8693 SMBsesssetupX (REQUEST) Flags1 0x10 (reserved) 0x08 (caseless) (Leach 3.2.2) Flags2 0x0 0x1 (KNOWS_LONG_NAMES) (Leach 3.2.3) Proc ID 28498 8593 UID 1 313 MID 2177 8693 Com2 0x75 0xFF Off2 116 0 MaxBuffer 2920 65535 MaxMpx 32 2 VcNumber 0 8593 (Illegal per Leach 4.1.2.1, see below) Capabilities 0x1 (not listed) 0x0 Account "GENFA" "LDOOLITT" Domain "" "WORKGROUP" OS "Windows 4" "Unix" LanMan "Windows 4" "Samba" SMBsesssetupX (REPLY) Error class 0x0 0x2 Error code 0x0 1 Proc ID 28498 8593 UID 2048 0 MID 2177 8693 Samba returned a VcNumber of 8593, when the NetApp said NumVcs=1 in its SMBnegprot (REPLY). This is a violation of Leach's section on "Pre NT LM 0.12". VcNumber is not mentioned in Leach's section on "NT LM 0.12", which is what both Windows 4 and Samba negotiated with the NetApp.

2 1

4.3.3 and y2000
by Mark Richards 23 Oct '98

23 Oct '98

Hello I'm running OnTap 4.3.3 (CIFS and NFS). I understand that the newest version of ontap is y2000 compliant. Does anyone know about the older versions? Regards Mark Richards

3 3

Re: (small?) feature request
by Tom Reingold 21 Oct '98

21 Oct '98

On Sun, 18 Oct 1998 13:05:48 EDT, Brian Tao <taob(a)risc.org> wrote: > Netapp could put up a "domestic build" and an "international > build" on now.netapp.com, and post up the usual disclaimers and domain > name checks, etc. to satisfy export restrictions. I'd love to ssh > into a Netapp and ditch rsh/telnet. Brian is right, netapp ought to be able to do it and ought to do it too. But you could try what I do. I have a "console" host with lots of serial ports, each connected to one of my servers' console ports. ssh into that host. Then the serial line from there to the netapp is secure. This workaround does not negate Brian's point.

6 5

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 October 1998