toasters

toasters@lists.teaparty.net

13520 discussions

RE: exchange mailbox level recovery using NetApp
by Rod Newcomb 13 Jan '06

13 Jan '06

We've been working with Netapp and iSCSI for about a year now. For the most part, when their software works, it works very well. But when it doesn't it's pretty bad. For example, we have a Clustered set of 920c's. We use Qlogic HBA's and GigE copper. We've had multiple issues of when the MS cluster fails over, we end up with multiple -1018 errors in our databases that need to be cleaned up. We've had three teams of consultants in to try to determine what is causing the problem, but to no avail. We're running 7.0.0.1 on the Filer (not a great release.. Were working on upgrading). Windows 2003 Enterprise SP1. Exchange Enterprise SP1. iSCSI initiator 1.06 QLOGIC 4010C HBA's. 2.0.0.5 We have a dedicated VLAN for our storage network, and no errors are showing on the HBA's. The last issue we had, we couldn't recover our snapshot (verified with eseutil as being good at the time of the backup) as it said the database was still corrupted after recovering. We ended up loosing about eight hours of email for 30 users. Also, beware that iSCSI is VERY expensive on space. To get guaranteed snapshots, you have to reserver 100% extra space of your lun. That's about 30% more than your average NFS file systems. We didn't understand that going into it, and grossly underestimated our storage space needs. We've had better luck with SQL, but still some issues. In the lab, Single Mailbox restore worked very well. It basically "virtually" mounts the message store and you select the information you want to bring back into your production database. Exchange isn't the wiser. I just wish we could get down to the bottom of our database issues involving iSCSI. The speed is incredible, and iSCSI is a great solution for our company on paper. In practice, it's been a little different experience. Rod Newcomb -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Leeds, Daniel Sent: Friday, January 13, 2006 4:02 PM To: Brian Parent; Netapp List Subject: Re: exchange mailbox level recovery using NetApp On 1/13/06 1:19 PM, "Brian Parent" <bparent(a)calvin.ucsd.edu> wrote: > We're considering using NetApp storage via iSCSI LUNs for Exchange > email databases, but are concerned that we'll run into trouble due to > some obscure/hidden Microsoft magic. > > I suspect that the way NetApp does snapshots, and presents the disk as > a LUN to the exchange server, it prevents any such problems. That is, > there is no way for the M$ software to know the difference between the > live production data, and the snapshot data from which you're making > mailbox level recovery. > > Anyone actually had success with such a setup? > Netapp makes a product called SnapManager for Exchange. Look into this product. It allows things such as single mailbox restores and more.

1 0

exchange mailbox level recovery using NetApp
by Brian Parent 13 Jan '06

13 Jan '06

We're considering using NetApp storage via iSCSI LUNs for Exchange email databases, but are concerned that we'll run into trouble due to some obscure/hidden Microsoft magic. I suspect that the way NetApp does snapshots, and presents the disk as a LUN to the exchange server, it prevents any such problems. That is, there is no way for the M$ software to know the difference between the live production data, and the snapshot data from which you're making mailbox level recovery. Anyone actually had success with such a setup?

3 2

RE: cifs terminate username?
by Ehrhart, Rick 13 Jan '06

13 Jan '06

Sorry about the internal URL. Here is the correct information: http://www.netapp.com/solutions/dfms/advantage.html This page talks about being an Advantage Developer. Just sign-up as an Advantage Developer in "Partner Level Desired". You can also talk to your sales person about making the SDK readily available for customers. - Rick - > -----Original Message----- > From: David Lee [mailto:t.d.lee@durham.ac.uk] > Sent: Friday, January 13, 2006 9:58 AM > To: Ehrhart, Rick > Subject: RE: cifs terminate username? > > > On Thu, 12 Jan 2006, Ehrhart, Rick wrote: > > > Your middle step might be easier if you used the > > cifs-session-list-iter-start, cifs-session-list-iter-next, and > > cifs-session-list-iter-end ONTAP APIs. NOW information is at: > > http://now.netapp.com/NOW/cgi-bin/manageontap. > Unfortunately, there is > > no cifs terminate API at this time, so you would still have > to use rsh > > to do the cifs terminate. > > (Off-list email...) > > Thanks. > > But I'm having trouble accessing that site. I can login (id: > netapplee ) > to the general NOW site, but cannot see: > http://now.netapp.com/NOW/cgi-bin/manageontap > > (it says "Unauthorized Access" etc.). > > > -- > > : David Lee I.T. Service : > : Senior Systems Programmer Computer Centre : > : Durham University : > : http://www.dur.ac.uk/t.d.lee/ South Road : > : Durham DH1 3LE : > : Phone: +44 191 334 2752 U.K. : >

2 1

RE: iSCSI to NetApp
by Holland, William L 13 Jan '06

13 Jan '06

How full is the aggregate/volume hosting the LUN? Could be WAFL performfance due to (nearly) full disks. In our development system I setup some several LUNs in different volumes on the same aggregate resulting in an aggregate that was 97% full -- to say the least, performance was horrible. Reconfigured to several smaller aggregates with no volume or aggregate > 90% full and performance went up dramatically. This is all using Microsoft's iSCSI initiator. You may also want to consider testing using software initiator vs hardware initiator. From what I've read and seen in some earlier testing I've done, the software initiator performed as well or better than the hardware initiator for MUCH less cost. -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:netbacker@gmail.com] Sent: Friday, January 13, 2006 1:41 PM To: David Lee Cc: toasters(a)mathworks.com Subject: Re: iSCSI to NetApp How big is the volume where the LUN resides? Is it on a seperate volume or on a heavily used one. Not enough spindles maybe? What model of filer? Could there be a routing/switching issue? Jumbo frames issue? I haven't used a TOE/iscsiHBA for a long time now, not worth the price. Standard Gig-E cards with Microsoft iSCSI drivers has been working great for us. Not much of a CPU hit. easiy to troubleshoot. If you have a spare NIC, why don't you try swapping the HBA with it and try MS iSCSI drivers to troubleshoot and check performance. -G On 1/13/06, David Lee <t.d.lee(a)durham.ac.uk> wrote: > (System: ONTAP 6.5.5) > > We have a Windows 2003 server (SP1) which uses iSCSI to mount a drive > onto a NetApps filer. > > When transfering files the performance is very poor: it can take 5 > minutes to transfer a 350 Mbyte file. Looking at Performance Monitor > on Windows while a file transfer is taking place shows that Disk Queue > Length readings can reach: > Average=94, Maximum=2142 > > We are using a QLogic QLA4010 Host Bus Adapter with the latest > firmware > (3.0.0.4) and driver (2.1.0.8). > > Any ideas? Thanks in advance. > > > -- > > : David Lee I.T. Service : > : Senior Systems Programmer Computer Centre : > : Durham University : > : http://www.dur.ac.uk/t.d.lee/ South Road : > : Durham DH1 3LE : > : Phone: +44 191 334 2752 U.K. : >

1 0

Message ("Your message dated Fri, 13 Jan 2006 10:46:26...")
by University of Arkansas LISTSERV Server (14.4) 13 Jan '06

13 Jan '06

Your message dated Fri, 13 Jan 2006 10:46:26 +0100 with subject "Mail System Error - Returned Mail" has been submitted to the moderator of the VMESA-L list: moderator(a)GIZMOWORKS.COM.

1 0

RE: cifs terminate username?
by Ehrhart, Rick 12 Jan '06

12 Jan '06

Your middle step might be easier if you used the cifs-session-list-iter-start, cifs-session-list-iter-next, and cifs-session-list-iter-end ONTAP APIs. NOW information is at: http://now.netapp.com/NOW/cgi-bin/manageontap. Unfortunately, there is no cifs terminate API at this time, so you would still have to use rsh to do the cifs terminate. - Rick - > -----Original Message----- > From: David Lee [mailto:t.d.lee@durham.ac.uk] > Sent: Thursday, January 12, 2006 3:56 AM > To: toasters(a)mathworks.com > Subject: cifs terminate username? > > > (System: ONTAP 6.5.5) > > >From the NetApp admin perspective, it would sometimes be > very useful to be > able to issue, via scripts, a command: > cifs terminate username > > Alas, this does not seem possible. The nearest seems to be > the sequence: > cifs sessions username > <human-read multiple-line output to extract a workstation > name: wsname> > cifs terminate wsname > > whose middle step is not script-friendly. > > > <background> > The underlying reason is that we sometimes change someone's username > (don't ask!). This requires us correspondingly to rename the final > component of their home directory. We do this from the UNIX > admin host, > (as "mv /path/to/oldname /path/to/newname") but this fails > if the user > has a CIFS session active. > </background> > > So an alternative to the above strategy would be a NetApp command that > could do the necessary rename of the directory. (There > doesn't seem to be > an "mv" or similar, although I do see an "mv" under > priv/advanced... but > I'm always slightly nervous around priv/advanced.) > > Any thoughts? > > -- > > : David Lee I.T. Service : > : Senior Systems Programmer Computer Centre : > : Durham University : > : http://www.dur.ac.uk/t.d.lee/ South Road : > : Durham DH1 3LE : > : Phone: +44 191 334 2752 U.K. : >

1 0

RE: cifs terminate username?
by Warkentin, Grant 12 Jan '06

12 Jan '06

FWW: We do user login ID renames from time to time ourselves. We don't do a lot of these so I have not considered scripting this. Our environment is Windows (Ontap 6.5.2R1) so this may not apply to you. If the user does have a CIFS session open I can terminate that session by using Microsoft's MMC to disconnect the user. We do warn them first that that this will happen. Basically, right click the my computer icon on my PC. Select manage. Right click computer management / connect to different computer and connect to the filer. Find the user's connection then disconnect it. Seems like a lot of work as the steps are written but I find it to be real simple & quick. I have no idea if it's possible to script this... Vbscript + WMI perhaps? NOTICE - This communication is intended ONLY for the use of the person or entity named above and may contain information that is confidential or legally privileged. If you are not the intended recipient named above or a person responsible for delivering messages or communications to the intended recipient, YOU ARE HEREBY NOTIFIED that any use, distribution, or copying of this communication or any of the information contained in it is strictly prohibited. If you have received this communication in error, please notify us immediately by telephone and then destroy or delete this communication, or return it to us by mail if requested by us. The City of Calgary thanks you for your attention and cooperation. -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of David Lee Sent: Thursday, January 12, 2006 4:56 AM To: toasters(a)mathworks.com Subject: cifs terminate username? (System: ONTAP 6.5.5) From the NetApp admin perspective, it would sometimes be very useful to be able to issue, via scripts, a command: cifs terminate username Alas, this does not seem possible. The nearest seems to be the sequence: cifs sessions username <human-read multiple-line output to extract a workstation name: wsname> cifs terminate wsname whose middle step is not script-friendly. <background> The underlying reason is that we sometimes change someone's username (don't ask!). This requires us correspondingly to rename the final component of their home directory. We do this from the UNIX admin host, (as "mv /path/to/oldname /path/to/newname") but this fails if the user has a CIFS session active. </background> So an alternative to the above strategy would be a NetApp command that could do the necessary rename of the directory. (There doesn't seem to be an "mv" or similar, although I do see an "mv" under priv/advanced... but I'm always slightly nervous around priv/advanced.) Any thoughts? -- : David Lee I.T. Service : : Senior Systems Programmer Computer Centre : : Durham University : : http://www.dur.ac.uk/t.d.lee/ South Road : : Durham DH1 3LE : : Phone: +44 191 334 2752 U.K. :

1 0

Re: cifs terminate username?
by Michael Bergman 12 Jan '06

12 Jan '06

>whose middle step is not script-friendly. Perhaps not, but it's IMO not so unfriendly as to make it *far* too expensive to parse correctly. So what I would try is to script it, using some powerful pattern matching tool(s) such as that available in Perl. I can see that it may take a rather skilled programmed to get this right. I agree with you that it would seem very convenient to be able to do cifs terminate <username> I suspect there's a potential problem here though, and it may be the reason for NetApp not having (yet) implemented that semantics: there may be several users connected from the same CIFS client (e.g. MS Windows host, or UNIX host with smbclient for example). If you then do 'cifs terminate <wsname>' it would not be the correct thing to do. I don't know if there would be 2 separate resources on the server side (in ONTAP CIFS handling) in such a case so that you can terminate one of them and leave the rest unaffected. If not, it's probably very hard to do a correct and logical implementaion of the semantics one wants here. Caveat lector: I'm not even close to knowing enough about CIFS inner workings to feel very sure about any of the above. >(There doesn't seem to be >an "mv" or similar, although I do see an "mv" under priv/advanced... but >I'm always slightly nervous around priv/advanced.) Does even that work if the directory is "held" by an actice CIFS session from a client? I suspect not. Also, you might not be able to script this, i.e. is it possible to do rsh <file> priv set advanced mv <from> <to> or do you have to be interactively connected to the console to use this mv? Cheers, M -- Michael Bergman Email: Michael.Bergman(a)ericsson.com Sr Systems Analyst EAB/OET/Z, R&D IT Stockholm Phone: +46 8 7275709 R&D IT Management Armborstv 14, S-125 25 Alvsjo, Sweden -- "Qui vicit non est victor nisi victus fatetur." - Ennius

1 0

cifs terminate username?
by David Lee 12 Jan '06

12 Jan '06

(System: ONTAP 6.5.5) From the NetApp admin perspective, it would sometimes be very useful to be able to issue, via scripts, a command: cifs terminate username Alas, this does not seem possible. The nearest seems to be the sequence: cifs sessions username <human-read multiple-line output to extract a workstation name: wsname> cifs terminate wsname whose middle step is not script-friendly. <background> The underlying reason is that we sometimes change someone's username (don't ask!). This requires us correspondingly to rename the final component of their home directory. We do this from the UNIX admin host, (as "mv /path/to/oldname /path/to/newname") but this fails if the user has a CIFS session active. </background> So an alternative to the above strategy would be a NetApp command that could do the necessary rename of the directory. (There doesn't seem to be an "mv" or similar, although I do see an "mv" under priv/advanced... but I'm always slightly nervous around priv/advanced.) Any thoughts? -- : David Lee I.T. Service : : Senior Systems Programmer Computer Centre : : Durham University : : http://www.dur.ac.uk/t.d.lee/ South Road : : Durham DH1 3LE : : Phone: +44 191 334 2752 U.K. :

1 0

New install
by Tim Longo 12 Jan '06

12 Jan '06

I'm in the process of setting up a new FAS270. The first thing I did was upgrade firmware/nd Ontap to version 7. After doing some reading, I think I will plan to implement FlexVols in the remaining free space. The root Vol0 seems to be bigger than I would have made it, and it seems it needs to be a traditional Vol that I cannot resize. I'd be interested to know what other people are doing here... are you leaving Vol0 as a traditional volume and adding other flex vols? What about resizing the root volume? Thanks.

4 4

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