toasters

toasters@lists.teaparty.net

3 participants
13514 discussions

Fabric MC and Brocade free ports usage
by Milazzo Giacomo 04 Dec '07

04 Dec '07

Hi All In a fabric MC enviroment a pair of Brocade 16 ports have been used for each site. Once filled with the needed connections to head and selve (Pool0 and Pool1) there are a lot of SFP empty with no connection. I know that other connection are not supported but I've already tried to attach a dual target HBA installed in each head and some HBA coming from hosts (Windows or Linux) and without zoning (you don't need it for MC) the Brocade I can expose all my FC LUNs created on NetApp and only to the right host using igroups with the 'authorized' WWWN. NetApp of course is in single_image cf mode. What do you think about this "unsupported" configuration? Many thanks

3 4

Can't expand SnapDrive for Windows drive
by Raj Patel 03 Dec '07

03 Dec '07

Just trying to expand a 20Gb lun used for SQL 2005 logs - Its contained within a 400Gb volume (I've allocated 140 for the database, 20 for logs and 10 for temp db). The vol is sized at well over twice the space requirements to allow for snaps. I just want to expand the log lun to 40Gb. Server is w2k3 sp1 SnapDrive 4.2.1 and OnTap 7.0.5 on a fas 270. SnapManager SQL is installed for SQL 2005. Overall SAN free space is about 80Gb. From within Windows the error SnapDrive gives is 'Unable to Expand Disk / Error there is not enough free space' Anything obvious I'm missing ? Thanks in advance, Raj.

2 1

Check Pointing
by Daryl Field 01 Dec '07

01 Dec '07

Hi Folks - we have some admittedly underpowered 270c in a mixed cifs/nfs environment. We are seeing a random performance issue whereby cp time is increased and the resultant nfs ops it serves drops considerably, resulting in an effective outage. The question is when I look at the sysstat output collected during these periods a lot of time is spent checkpointing 'special files' - its been a while since I covered ontap theory so if someone could explain this in further detail it would be helpful. The now docs omit extended explanations. Below is a sample of normal vs check pointing sysstat output. Thanks, Daryl Normal CPU NFS CIFS HTTP Total Net kB/s Disk kB/s Tape kB/s Cache Cache CP CP Disk FCP iSCSI FCP kB/s in out read write read write age hit time ty util in out 92% 1758 2 0 1760 495 2621 6404 0 0 0 41s 75% 0% - 55% 0 0 0 0 89% 2759 3 0 2762 781 6441 11140 6 0 0 22s 83% 0% - 68% 0 0 0 0 96% 2605 20 0 2625 1566 7347 12961 24 0 0 41s 78% 0% - 70% 0 0 0 0 97% 2224 3 0 2227 688 4856 8198 0 0 0 41s 84% 0% - 49% 0 0 0 0 92% 4027 2 0 4029 1860 5267 9829 12 0 0 14s 89% 0% - 72% 0 0 0 0 96% 2136 0 0 2136 535 1809 6405 17 0 0 42s 85% 0% - 64% 0 0 0 0 95% 2383 0 0 2383 642 3245 8769 0 0 0 42s 82% 0% - 67% 0 0 0 0 97% 2210 0 0 2210 595 3823 8309 1578 0 0 14s 89% 66% Tn 61% 0 0 0 0 85% 4889 0 0 4889 1881 5163 15287 7252 0 0 23s 85% 100% :s 100% 0 0 0 0 90% 2393 3 0 2396 817 10048 18018 2474 0 0 39s 89% 100% :s 70% 0 0 0 0 95% 1603 18 0 1621 913 4532 8437 4025 0 0 40s 88% 100% :v 54% 0 0 0 0 94% 2260 0 0 2260 1109 4061 8470 51 0 0 40s 82% 12% : 61% 0 0 0 0 88% 3532 3 1 3536 890 5716 11892 0 0 0 24s 83% 0% - 71% 0 0 0 0 92% 3347 2 0 3349 2197 6462 11603 20 0 0 23s 87% 0% - 68% 0 0 0 0 94% 3607 17 0 3624 1013 6285 11300 7 0 0 15s 85% 0% - 69% 0 0 0 0 95% 2359 7 0 2366 598 4574 8985 0 0 0 41s 84% 0% - 69% 0 0 0 0 94% 3981 0 0 3981 1098 4065 10324 23 0 0 39s 86% 0% - 74% 0 0 0 0 83% 3965 0 0 3965 946 4709 11296 0 0 0 39s 82% 0% - 83% 0 0 0 0 88% 3318 0 0 3318 1009 5220 11772 24 0 0 23s 83% 0% - 86% 0 0 0 0 88% 2821 0 0 2821 750 3974 9762 8 0 0 14s 79% 0% - 69% 0 0 0 0 Check Pointed CPU NFS CIFS HTTP Total Net kB/s Disk kB/s Tape kB/s Cache Cache CP CP Disk FCP iSCSI FCP kB/s in out read write read write age hit time ty util in out 98% 1651 7 0 1658 669 2196 8671 15 0 0 49s 88% 0% - 44% 0 0 0 0 99% 567 0 0 567 176 1022 2601 3 0 0 46s 87% 0% - 41% 0 0 0 0 99% 614 0 0 614 205 1632 3408 8 0 0 1 89% 22% :n 27% 0 0 0 0 99% 929 1 0 930 252 1803 3865 237 0 0 44s 89% 100% Tn 32% 0 0 0 0 98% 463 0 0 463 385 1338 4066 1224 0 0 1 91% 100% :n 32% 0 0 0 0 98% 398 6 0 404 145 1894 5134 1787 0 0 1 81% 100% :s 34% 0 0 0 0 99% 558 0 0 558 186 1716 3675 912 0 0 50s 89% 100% :s 34% 0 0 0 0 99% 677 1 0 678 919 902 2892 587 0 0 1 87% 100% :s 33% 0 0 0 0 99% 703 1 0 704 293 1179 3173 731 0 0 52s 89% 100% :s 26% 0 0 0 0 97% 2250 17 0 2267 717 3024 10358 3462 0 0 52s 94% 100% :v 63% 0 0 0 0 98% 1032 0 0 1032 619 933 2482 81 0 0 49s 92% 11% : 33% 0 0 0 0 99% 875 1 0 876 293 2328 3818 6 0 0 44s 92% 0% - 33% 0 0 0 0 99% 325 0 0 325 116 579 1595 5 0 0 42s 91% 0% - 23% 0 0 0 0 99% 884 1 0 885 382 1613 3723 10 0 0 1 93% 38% :n 24% 0 0 0 0 98% 917 0 0 917 586 1193 3136 415 0 0 42s 88% 100% Tn 31% 0 0 0 0 99% 675 0 0 675 462 554 4040 2610 0 0 1 90% 100% :s 28% 0 0 0 0 98% 718 0 0 718 396 990 2469 355 0 0 44s 92% 100% :s 30% 0 0 0 0 99% 555 1 0 556 181 965 2101 257 0 0 1 89% 100% :s 33% 0 0 0 0 99% 824 0 0 824 391 1268 3484 877 0 0 48s 97% 100% :s 29% 0 0 0 0 97% 2119 14 0 2133 2154 2389 5898 347 0 0 45s 93% 100% :f 50% 0 0 0 0

8 9

RE: symbolic links / wide links
by Kevin Parker 30 Nov '07

30 Nov '07

Have implemented these in an 5.5TB and 8500+ user environment. No issues whatsoever, worked like a champ, stick with Ontap 6.5 or above. Enable strict symlink option from the cifs shares command. Plan on using either 1) a *nix client to mount file the filesystem 2) a windows client that can talk *nix file protocols (SFU does not work great, try cygwin or others) 3) search NOW for "symlink tool", from NetApp's Toolbox. Very slick. In fact, use this option first if u r a windows shop. **Sent using wireless handheld... please excuse any typo's or brevity** Kevin Parker -----Original Message----- From: "Joel.Cooper(a)lvvwd.com" <Joel.Cooper(a)lvvwd.com> To: "owner-toasters(a)mathworks.com" <owner-toasters(a)mathworks.com>; "toasters(a)mathworks.com" <toasters(a)mathworks.com> Sent: 11/30/2007 12:15 PM Subject: symbolic links / wide links Is anyone using symbolic links/widelinks with their filer in a cifs environment? Have you experienced any problems? Any suggestions? ________________________________ Note: This message and any attachments is intended solely for the use of the individual or entity to which it is addressed and may contain information that is non-public, proprietary, legally privileged, confidential, and/or exempt from disclosure. If you are not the intended recipient, you are hereby notified that any use, dissemination, distribution, or copying of this communication is strictly prohibited. If you have received this communication in error, please notify the original sender immediately by telephone or return email and destroy or delete this message along with any attachments immediately.

1 0

Re: Expand LUN for Exch data
by Bill Holland 30 Nov '07

30 Nov '07

It's a simple one-step process within Snapdrive and results in no downtime. I'm pretty sure that growing the LUN, outside of snapdrive, while it is mounted will confuse Windows and result in disruption (and therefore downtime) to Exchange. ----- Original Message ----- From: "Bill Holland" <hollandwl(a)gmail.com> To: "Kevin Parker" <KParker(a)nwnit.com>; "NetApp list" <toasters(a)mathworks.com> Sent: Friday, November 30, 2007 10:04 AM Subject: Re: Expand LUN for Exch data > This is possible with Snapdrive, not possible without it. > > ----- Original Message ----- > From: "Kevin Parker" <KParker(a)nwnit.com> > To: "NetApp list" <toasters(a)mathworks.com> > Sent: Friday, November 30, 2007 9:05 AM > Subject: Expand LUN for Exch data > > >> I'm off network so I can't research this, but want to ensure that I can >> expand a LUN for Exchange on the fly, no downtime. >> Plan to use SnapDrive 5.0. Volume is already large enough in capacity. >> >> Thanks!! >> >> Best regards, >> Kevin M. Parker >> >> Note: This message and any attachments is intended solely for the use of >> the individual or entity to which it is addressed and may contain >> information that is non-public, proprietary, legally privileged, >> confidential, and/or exempt from disclosure. If you are not the intended >> recipient, you are hereby notified that any use, dissemination, >> distribution, or copying of this communication is strictly prohibited. >> If you have received this communication in error, please notify the >> original sender immediately by telephone or return email and destroy or >> delete this message along with any attachments immediately. >> >

2 1

Can you allocate space from one aggregate to a volume on another?
by Bryano 29 Nov '07

29 Nov '07

We currently have a fas 940 with the following config: 3x 144GB FC shelves all but 2 disks are in aggr1 which equates to approx 3.9TB useable space which is approx 5.7TB raw (excluding the 2 spares) containing 4 volumes 5x 72GB FC shelves with 8 assigned to aggr2 (60 spares) containing 1 volume We would like to increase the size of aggr1 and therefore plan to repalce 2 of the 72GB shelves with 144Gb, 4 of which will be retained as spares, thus giving us an extra 3.4TB of raw space to add to aggr1. Looking at the system configuration guide for the fas940 it states that the max raw aggregate size is 8TB therefore this would take aggr1 over the limit. Is it possible to create a new aggregate (aggr3) assign the new 144GB drives to it but assign the space to the volumes held on aggr1? As it stands if we were to replace 2 shelves with 144GB drives and remove 1 72GB shelf we would have a max size (based on 5x 144GB and 2x 72GB shelves) of 12TB which takes us to the max raw capacity of this filer using FC-AL drives. Thanks -- View this message in context: http://www.nabble.com/Can-you-allocate-space-from-one-aggregate-to-a-volume… Sent from the Network Appliance - Toasters mailing list archive at Nabble.com.

11 15

iSCSI question
by David McWilliams 29 Nov '07

29 Nov '07

Is there a requirement for the volume that contains an iSCSI LUN to have a minimum amount of free space. We seem to be getting warnings when it approaches 90% full. Is it a modifiable parameter? -- Sláinte, David Get a safer, faster, better web browser @ http://www.mozilla.org/products/firefox/

4 4

Re: Can you allocate space from one aggregate to a volume on another?
by Robert Lazzurs 23 Nov '07

23 Nov '07

Widelinks look very cool, sadly the only data on this volume are LUN's. On Nov 23, 2007 2:01 PM, Agnew, Blaine <Blaine.Agnew(a)netapp.com> wrote: > Have you checked out widelinks? > > In certain circumstances I have used these to make data appear to be in > one volume when in fact it existed in others. I think you would need NFS > enabled in order to use this conveniently. > > Blaine > > > -----Original Message----- > From: Holland, William L [mailto:HollandWL@state.gov] > Sent: Friday, November 23, 2007 7:52 AM > To: Robert Lazzurs; toasters(a)mathworks.com > Subject: RE: Can you allocate space from one aggregate to a volume on > another? > > You can use snapmirror to migrate the volume to a new aggregate. When > you are ready to make the final cutover, rename the original volume then > rename the new volume to match the original volume name. You will also > need to move your CIFS shares (if you have any). Then you can grow the > new volume to the size you need. This will be a minimal disruption to > users and transparent as far as their applications, etc are concerned. > > -----Original Message----- > From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] > On Behalf Of Robert Lazzurs > Sent: Friday, November 23, 2007 6:55 AM > To: toasters(a)mathworks.com > Subject: Re: Can you allocate space from one aggregate to a volume on > another? > > Is there any way this can be done that is fully transparent to the end > users of the storage. > > We have a similar situation and while we will be moving to GX in the > long term this is not something that can be done in the short term and > we are fast running out of space on our original aggr. > > If a volume cannot be moved to another aggr without the end users > noticing can a LUN be moved to another volume without end users > noticing? > > Kind regards, > > On Nov 15, 2007 4:48 PM, Fox, Adam <Adam.Fox(a)netapp.com> wrote: > > If I understand you correctly, I don't think you can do what you want > > to do. > > It sounds like you want a volume to have space allocated in multiple > > aggregates. > > That won't work. You could copy an entire volume into a new > > aggregate, but you can't have a flex vol span aggregates. At least > > not in ONTAP 7. ONTAP GX can do this, but that's a very different > > configuration and you'd need to work closely with your sales team to > > see if it's appropriate for your environment. > > > > > > -- Adam Fox > > adamfox(a)netapp.com > > > > > > > > -----Original Message----- > > From: Bryano [mailto:Bryan.Price@firstassistinsurance.co.uk] > > Sent: Thursday, November 15, 2007 7:48 AM > > To: toasters(a)mathworks.com > > Subject: Can you allocate space from one aggregate to a volume on > > another? > > > > > > We currently have a fas 940 with the following config: > > > > 3x 144GB FC shelves all but 2 disks are in aggr1 which equates to > > approx 3.9TB useable space which is approx 5.7TB raw (excluding the 2 > > spares) containing 4 volumes > > > > 5x 72GB FC shelves with 8 assigned to aggr2 (60 spares) containing 1 > > volume > > > > We would like to increase the size of aggr1 and therefore plan to > > repalce 2 of the 72GB shelves with 144Gb, 4 of which will be retained > > as spares, thus giving us an extra 3.4TB of raw space to add to aggr1. > > Looking at the system configuration guide for the fas940 it states > > that the max raw aggregate size is 8TB therefore this would take aggr1 > > > over the limit. > > > > Is it possible to create a new aggregate (aggr3) assign the new 144GB > > drives to it but assign the space to the volumes held on aggr1? > > > > As it stands if we were to replace 2 shelves with 144GB drives and > > remove 1 72GB shelf we would have a max size (based on 5x 144GB and 2x > > > 72GB shelves) of 12TB which takes us to the max raw capacity of this > > filer using FC-AL drives. > > > > Thanks > > > > > > -- > > View this message in context: > > http://www.nabble.com/Can-you-allocate-space-from-one-aggregate-to-a-v > > ol > > ume-on-another--tf4811616.html#a13766859 > > Sent from the Network Appliance - Toasters mailing list archive at > > Nabble.com. > > > > >

1 0

san redesign and configuration questions
by Alexander Schalek 23 Nov '07

23 Nov '07

Hi everyone, I have some questions for the experts regarding a SAN configuration. Configuration: FAS3020c MetroCluster (2 sites, 16km distance) 2 FC-Switches on each site, each connected via ISL's to the other location. 1 RHEL ES4 U5 Host on each site with a 4GB QLogic Dual Port HBA. Each FC-Port on each Host is connected to one of the switches FCP enabled and the cf mode is single image Snapdrive for Linux on both Hosts. Following situation: Snapdrive for Linux currently does not support DM-Multipathing on RedHat Linux. We have an open case with Netapp and they are working on the issue to support mulipathing (planned for Q1/08) When I started I setup everything for multipathing because we thought the support for it would be released earlier. I zoned the switches... Zoning: switch1 zone1: 0b(filer1)+0b(filer2)+port0(host1) switch1 zone2: 0b(filer1)+0b(filer2)+port1(host2) switch2 zone1: 0d(filer1)+0d(filer2)+port1(host1) switch2 zone2: 0d(filer1)+0d(filer2)+port0(host2) On the host I implemented DM-MP and use the mpath-Device's to access the LUN's on the cluster. Each mpath-Device consists of 4 sdx-Devices(=the 4 paths to the LUN). Example: mpath0 (360a98000433464706334455142706e33) [size=100 MB][features="1 queue_if_no_path"][hwhandler="0"] \_ round-robin 0 [enabled] \_ 3:0:1:0 sdd 8:48 [active][ready] \_ 4:0:1:0 sdf 8:80 [active][ready] \_ round-robin 0 [enabled] \_ 3:0:0:0 sdc 8:32 [active][ready] \_ 4:0:0:0 sde 8:64 [active][ready] This is working fine. The customer already installed his Oracle RAC on the LUN's/mpath-Device's which where presented to the host's. And here my problem: The customer can not wait until Q1/08 when Snapdrive for Linux supports multipathing on RHEL and I have to redesign the SAN to a single path configuration so SnapDrive is working and supported. My questions: Can I just use the sdx-Devices on the linux host's and leave the zoning and multipathing as it is? Will the cluster failover then work correctly? Or should I deactive multipathing on the hosts and just zone one hba port to one filer fc port and use that sdx-Device? I will have to implement multipathing for sure later when it is supported with linux and snapdrive so I prefer minimum or no changes to the current configuration :) Has anyone experience with such a problem/configuration? Thankful for any tips & hints! Best regards, Alexander -- Alexander Schalek, Technical Services / IT-Infrastruktur Bacher Systems EDV GmbH FN 54202i, Handelsgericht Wien Clemens-Holzmeister-Strasse 4 A-1100 Wien, Business Park Vienna, Austria phone: +43 (1) 60 126-34 | fax: +43 (1) 60 126-4 e-mail: as(a)bacher.at<mailto:as@bacher.at> | web: www.bacher.at<http://www.bacher.at>

1 0

FW Generic Failure
by Michail Michalakis 22 Nov '07

22 Nov '07

Mon Nov 19 12:56:46 EET [esh.bypass.host.port:CRITICAL]: replayed event: [IN] port on channels 0b/UNUSED shelf id 2 ESH B Bypassed due to FW generic fault Does any understand the above message is was found on a FAS270, an the aggregate on this volumes status is online, raid-dp, redirect. I suspect that it is the FAS270 controller and not the ESH Module, because when changing the modules the error remains. All LED on front and back are green.

4 3

Jump to page:

2025

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