toasters

toasters@lists.teaparty.net

2 participants
13517 discussions

Home directories become owned by root
by Michael Coxe 04 May '05

04 May '05

Hi, We have a Netapp using shared NFS/CIFS. Yesterday several accounts became owned by root. Users couldn't even cd to their homes via NFS, though from the Windows side all looked OK. As root from the administration server I couldn't access their directories either. I thought of restoring from .snapshot but those directories (nightly.o, etc) are also onwed by root and are inaccessable from the Unix side. The only way to allow access was to use SecureShare from the Windows side and change ownership from "root" back to the username. Files & subdirectories appear to be owned by the user but the Unix permissions are all set to 777, and they can't access these files/directories. I could reset them via SecureShare (and did so for 2 users) but they won't be what the user(s) originally set them to. Ontap: NetApp Release 6.1.2D1 Filer: F740 Anyone have an idea how I can restore the original ownership & file permissions other that restoring from tape. -- Michael Coxe, IT <mcoxe(a)opsware.com>

1 0

RE: Simple Question on QTREES and Volumes
by Warkentin, Grant 04 May '05

04 May '05

See Below: 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 Paul Jordison (AU) Sent: Wednesday, May 04, 2005 12:31 AM To: toasters(a)mathworks.com Subject: Simple Question on QTREES and Volumes Hi, I will preface this by stating that I am very green when it comes to NetApp things. I was in the same position last year. Still am to a certain extent. http://now.netapp.com is your friend! OK, we have a Filer (not sure of the exact model...I can post that if needed) and I am not sure of the level of software we use (once again I can find this) but my question is about Disks, Volumes and Qtrees. The Model does not matter too much. But the version of DataOnTap is important. We're currently running version 6.5x. Version 7 brings extra functionality (flex Vols). A volume is a bunch of disks (LUNS) yes? - ie if I get 5 66GB disks RAID 4 it gives me about a 200GB volume? - The size of the volume is directly proportional to the amount of disks added to it?? Yes, a volume is just a bunch of disks. Qtrees are effectively virtual volumes that 'break up' the volume even more (ie a 200GB volume can have 8 x 25GB Qtrees (or 1 x 50GB Qtree and 6 x 25GB Qtrees)? No, Think of a qTree as a folder on a disk. You can assign quota to a qTree to control how much it will grow. 5 qtrees on a 500 meg volume = each qtree can store a max of 500 meg. Or a combination of 500 meg till the disk gets full. Qtrees can be increased or decreased at any level (if I add a disk to a volume all of that space can go to a Qtree or none of it while the volume MUST increase by the amount of physical disk added? Only adding disks to the volume will affect the total space available to all qtrees. Space assigned to qtrees is controlled by quotas. If I want to put data on the filer (say Exchange data, user homedrives and Departmental data) I can put this data on separate Qtrees in a single volume? Or should I put the data on separate volumes and not worry about Qtrees? It's your choice. We have four volumes currently. One for livelink data One for UNIX data Two for the NT Stuff. Putting a quota on home drive was one of my directives so I have 5 qtrees dedicated to home drives. Two trees with a default quota of 250 meg / user. One with a 500m / user, one @ 1 gig / user, the last one is a 2 gig / user default plus the folders of those that go over 2 gig (quotas have to be applied manually for the over 2 gig users). The rest of the space on the NT volumes are used for departmental shares. These shares reside in their own qtrees. These qtrees have tree quotas on them (mainly to limit growth). Is it better to have 1 massive volume and split it into qtrees as needed or is there a limit on volume size (performance problems if volume too big?)? There is a limit of 250 qtrees per volume if I recall correctly. Any folder you put in the root of a volume is considered a Qtree. Sorry if this is totally incorrect...if it is please feel free to correct me as I said I have had nothing to do with NetApp stuff before (although I am quite handy on a HDS SAN :)) But if I am right, please let me know as I need this information. Sorry for taking so long with this email. HTH if I can. Grant Warkentin System Administrator II City of Calgary - Information Technology Mail Drop == 8245 Ph: (403) 268-5629 Fax: (403) 268-2546

1 0

Simple Question on QTREES and Volumes
by Paul Jordison (AU) 04 May '05

04 May '05

Hi, I will preface this by stating that I am very green when it comes to NetApp things. OK, we have a Filer (not sure of the exact model...I can post that if needed) and I am not sure of the level of software we use (once again I can find this) but my question is about Disks, Volumes and Qtrees. A volume is a bunch of disks (LUNS) yes? - ie if I get 5 66GB disks RAID 4 it gives me about a 200GB volume? - The size of the volume is directly proportional to the amount of disks added to it?? Qtrees are effectively virtual volumes that 'break up' the volume even more (ie a 200GB volume can have 8 x 25GB Qtrees (or 1 x 50GB Qtree and 6 x 25GB Qtrees)? Qtrees can be increased or decreased at any level (if I add a disk to a volume all of that space can go to a Qtree or none of it while the volume MUST increase by the amount of physical disk added? If I want to put data on the filer (say Exchange data, user homedrives and Departmental data) I can put this data on separate Qtrees in a single volume? Or should I put the data on separate volumes and not worry about Qtrees? Is it better to have 1 massive volume and split it into qtrees as needed or is there a limit on volume size (performance problems if volume too big?)? Sorry if this is totally incorrect...if it is please feel free to correct me as I said I have had nothing to do with NetApp stuff before (although I am quite handy on a HDS SAN :)) But if I am right, please let me know as I need this information. Sorry for taking so long with this email.

1 0

RE: VFM and DFS
by Fahey_Michael＠emc.com 03 May '05

03 May '05

Guys, Let's stop vendor bashing in public. It is not good form and detracts from the spirit of the board. Everyone have a nice day. Mike -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Eric Kimminau Sent: Monday, May 02, 2005 8:14 PM To: toasters(a)mathworks.com Subject: VFM and DFS VFM or NuView StorageX (they are the exact same product) is an enhanced management GUI for Microsoft DFS with some additional hooks for automating failover in a NetApp (and other?) environment. I have had a number of conversations with the Microsoft DFS project manager. If there is anything Microsoft hates more than anything else, it is telling a customer that they need to purchase a third party product to manage a Microsoft service. This means that either A) Microsoft is going to purchase technology so that they can give it to you for free or B) they are going to develop technology so that they can give it to you for free. Either way a commercial third party product for managing DFS has a relatively short shelf life with a fairly steep price tag if you have significant amount of storage. NuView knows this and has started bundling additional "modules" with their StorageX/VFM product to make it a more attractive solution. All I can say is that you need to take a close look at where it has been installed and ask to talk to customers that have implemented the add-on modules to see what level of success they have seen. If you are considering a 3rd party tool to configure and manage your DFS environment you should understand a couple of things. 1) You MUST be running MS Server 2003 for your DFS root. Win200 is missing some key functionality and stability enhancements which will create a less than stable and less than scaleable DFS environment. Even VFM/StorageX require a Windows 2003 Server./ Their GUI is just that, a "better" GUI for administering DFS. It still requires MS Windows DFS. 2) Your DFS root will either need a Active Directory Domain Controller(s) or a dedicated DFS root server (cluster) depending upon the size and complexity of your environment and the level of authority you want to assign to your DFS administrators. 3) If you want to delegate the administration of DFS to less than domain administrators, you cannot install your DFS root on your AD domain controllers. You will be implementing distributed domain DFS. This means you need one or more clusters of DFS root servers to insure the availability of your DFS resources. 4) If you install DFS on your domain controllers, you must create a domain administrator account to administer DFS. 5) DFS R2 is currently in beta. Patches have been released in December and again in 2005 (March?) which have significantly improved the Microsoft tools for configuring and managing DFS. If you are seriously looking gat DFS, contact Microsoft and ask them to come in and show you what they have currently available for free before you spend any significant time or $$ or a tool you will have to pay for. If what you are really looking for is a tool to transparently monitor, manage and optimize your storage environment, both CIFS and NFS and support multiple vendors without having to install software on your filers or clients, without having to install Windows servers and 3rd party software to manage your heterogeneous environment, there are significantly more robust solutions which integrate into industry standard global name spaces and which ELIMINATE downtime, not just reduce it. YMMV. My $1.25. E. -- ---------1---------2---------3---------4---------5---------6---------7 Eric Kimminau

3 2

VFM and DFS
by Eric Kimminau 03 May '05

03 May '05

VFM or NuView StorageX (they are the exact same product) is an enhanced management GUI for Microsoft DFS with some additional hooks for automating failover in a NetApp (and other?) environment. I have had a number of conversations with the Microsoft DFS project manager. If there is anything Microsoft hates more than anything else, it is telling a customer that they need to purchase a third party product to manage a Microsoft service. This means that either A) Microsoft is going to purchase technology so that they can give it to you for free or B) they are going to develop technology so that they can give it to you for free. Either way a commercial third party product for managing DFS has a relatively short shelf life with a fairly steep price tag if you have significant amount of storage. NuView knows this and has started bundling additional "modules" with their StorageX/VFM product to make it a more attractive solution. All I can say is that you need to take a close look at where it has been installed and ask to talk to customers that have implemented the add-on modules to see what level of success they have seen. If you are considering a 3rd party tool to configure and manage your DFS environment you should understand a couple of things. 1) You MUST be running MS Server 2003 for your DFS root. Win200 is missing some key functionality and stability enhancements which will create a less than stable and less than scaleable DFS environment. Even VFM/StorageX require a Windows 2003 Server./ Their GUI is just that, a "better" GUI for administering DFS. It still requires MS Windows DFS. 2) Your DFS root will either need a Active Directory Domain Controller(s) or a dedicated DFS root server (cluster) depending upon the size and complexity of your environment and the level of authority you want to assign to your DFS administrators. 3) If you want to delegate the administration of DFS to less than domain administrators, you cannot install your DFS root on your AD domain controllers. You will be implementing distributed domain DFS. This means you need one or more clusters of DFS root servers to insure the availability of your DFS resources. 4) If you install DFS on your domain controllers, you must create a domain administrator account to administer DFS. 5) DFS R2 is currently in beta. Patches have been released in December and again in 2005 (March?) which have significantly improved the Microsoft tools for configuring and managing DFS. If you are seriously looking gat DFS, contact Microsoft and ask them to come in and show you what they have currently available for free before you spend any significant time or $$ or a tool you will have to pay for. If what you are really looking for is a tool to transparently monitor, manage and optimize your storage environment, both CIFS and NFS and support multiple vendors without having to install software on your filers or clients, without having to install Windows servers and 3rd party software to manage your heterogeneous environment, there are significantly more robust solutions which integrate into industry standard global name spaces and which ELIMINATE downtime, not just reduce it. YMMV. My $1.25. E. -- ---------1---------2---------3---------4---------5---------6---------7 Eric Kimminau

1 0

test - is list still alive
by Dan O'Brien 02 May '05

02 May '05

test - no need to respond -- Dan O'Brien, dmobrien(a)lcsi.net Cell: 614-783-4859 Work: 614-476-8473

2 3

RE: clear arp on Netapp?
by Cao, Vinh V (Vinh) 02 May '05

02 May '05

Hey Venkat, That works. Thank you very much. Vinh > -----Original Message----- > From: Venkat Appineni [mailto:appinenv@yahoo-inc.com] > Sent: Monday, May 02, 2005 2:28 PM > To: Cao, Vinh V (Vinh); toasters(a)mathworks.com > Subject: RE: clear arp on Netapp? > > > Vinh, > > You can avoid a reboot by doing the following > > Source /etc/rc > > This should reread the /etc/rc file and build the new arp table.. > > -Venkat > > -----Original Message----- > From: owner-toasters(a)mathworks.com > [mailto:owner-toasters@mathworks.com] > On Behalf Of Cao, Vinh V (Vinh) > Sent: Monday, May 02, 2005 10:51 AM > To: 'toasters(a)mathworks.com' > Subject: clear arp on Netapp? > > Hello All, > > Does Data Ontapp allow clear arp catch? > > I have F820 Ontapp Version 6.4.2P6. I just need > to clear arp. However, I don't see the command that > is allow me to do clear arp catch. > I need clear arp, not just do "apr -d host_name" > > Here is the problem I have: On filer view I see two host > that are in arp tables, but when I telnet to my file > I do arp -a, it does not show up. > > I tried arp -d host_name, then I got error: > > arp: delete: cannot locate patriots ( patriots is Unix client that map > to filer ) > > From filer view: > > snip ................. > > default router UGS 48 7829558 e4 > routerv3 0:d0:bc:f5:2:f8 UHL 0 0 e10-3 > xena 0:6:5b:3c:e:b0 UHL 0 1 e10-3 > hercules 0:6:5b:3c:e:5c UHL 0 2465 e10-3 > 10.3.101.241 0:b0:d0:c3:1b:2f UHL 0 0 e10-3 > 10.3.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-3 > subtract pix UGHD 0 0 e4 > 10.9/16 link#8 UC 0 0 e10-9 > routerv9 0:d0:bc:f5:2:f8 UHL 0 0 e10-9 > 10.9.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-9 > patriots pix UGHD 0 13512 e4 > localhost localhost UH 13 1805 lo > > snip .................. > > Clearly I want to remove "subtract" and "patriots" from routing table. > These two are right now not able to map to my filer. > > Does any one know? or the only way is reboot? > > Thanks, > > Vinh >

1 0

RE: clear arp on Netapp?
by Venkat Appineni 02 May '05

02 May '05

Vinh, You can avoid a reboot by doing the following Source /etc/rc This should reread the /etc/rc file and build the new arp table.. -Venkat -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Cao, Vinh V (Vinh) Sent: Monday, May 02, 2005 10:51 AM To: 'toasters(a)mathworks.com' Subject: clear arp on Netapp? Hello All, Does Data Ontapp allow clear arp catch? I have F820 Ontapp Version 6.4.2P6. I just need to clear arp. However, I don't see the command that is allow me to do clear arp catch. I need clear arp, not just do "apr -d host_name" Here is the problem I have: On filer view I see two host that are in arp tables, but when I telnet to my file I do arp -a, it does not show up. I tried arp -d host_name, then I got error: arp: delete: cannot locate patriots ( patriots is Unix client that map to filer ) From filer view: snip ................. default router UGS 48 7829558 e4 routerv3 0:d0:bc:f5:2:f8 UHL 0 0 e10-3 xena 0:6:5b:3c:e:b0 UHL 0 1 e10-3 hercules 0:6:5b:3c:e:5c UHL 0 2465 e10-3 10.3.101.241 0:b0:d0:c3:1b:2f UHL 0 0 e10-3 10.3.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-3 subtract pix UGHD 0 0 e4 10.9/16 link#8 UC 0 0 e10-9 routerv9 0:d0:bc:f5:2:f8 UHL 0 0 e10-9 10.9.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-9 patriots pix UGHD 0 13512 e4 localhost localhost UH 13 1805 lo snip .................. Clearly I want to remove "subtract" and "patriots" from routing table. These two are right now not able to map to my filer. Does any one know? or the only way is reboot? Thanks, Vinh

1 0

clear arp on Netapp?
by Cao, Vinh V (Vinh) 02 May '05

02 May '05

Hello All, Does Data Ontapp allow clear arp catch? I have F820 Ontapp Version 6.4.2P6. I just need to clear arp. However, I don't see the command that is allow me to do clear arp catch. I need clear arp, not just do "apr -d host_name" Here is the problem I have: On filer view I see two host that are in arp tables, but when I telnet to my file I do arp -a, it does not show up. I tried arp -d host_name, then I got error: arp: delete: cannot locate patriots ( patriots is Unix client that map to filer ) From filer view: snip ................. default router UGS 48 7829558 e4 routerv3 0:d0:bc:f5:2:f8 UHL 0 0 e10-3 xena 0:6:5b:3c:e:b0 UHL 0 1 e10-3 hercules 0:6:5b:3c:e:5c UHL 0 2465 e10-3 10.3.101.241 0:b0:d0:c3:1b:2f UHL 0 0 e10-3 10.3.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-3 subtract pix UGHD 0 0 e4 10.9/16 link#8 UC 0 0 e10-9 routerv9 0:d0:bc:f5:2:f8 UHL 0 0 e10-9 10.9.255.255 ff:ff:ff:ff:ff:ff UHL 0 53675 e10-9 patriots pix UGHD 0 13512 e4 localhost localhost UH 13 1805 lo snip .................. Clearly I want to remove "subtract" and "patriots" from routing table. These two are right now not able to map to my filer. Does any one know? or the only way is reboot? Thanks, Vinh

1 0

de-subscribe
by Gordy Peck 02 May '05

02 May '05

1 0

RE: SMB tuning/understanding
by Alan Biren 29 Apr '05

29 Apr '05

OpLkBkDelayedBreak is when an oplock break is sent too soon after an open request completes, some clients lock up. This variable counts the number of times that Data ONTAP must delay before sending an oplock break. The delay is controlled by the cifs.oplocks.opendelta option. In case you need a definition of oplock here it is: Oplocks (opportunistic locks) enable a CIFS client in certain file-sharing scenarios to perform client-side caching of read-ahead, write-behind, and lock information. A client can then read from or write to a file without regularly reminding the server that it needs access to the file in question. This improves performance by reducing network traffic. The NOW site has some fairly simplistic documentation into CIFS Performance Tuning. Alan -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Simon Waters Sent: Friday, April 29, 2005 6:04 AM To: toasters(a)mathworks.com Subject: SMB tuning/understanding Pondering some Microsoft tuning guidelines in one window, and the Filer report in another. What is "OpLkBkDelayedBreak"? And better yet where do I find an idiots guide to the innards of Microsofts networking? ***************************************************************************************************** If you have received this email in error please notify the the system manager or the sender immediately and do not disclosed the to anyone or make copies. ** eSafe scanned this email for viruses, vandals or malicious content ** ** To find out eSafe(TM) and other Aladdin Security Products, contact AMC Corporation at (212) 736-0111 *****************************************************************************************************

1 0

RE: VFM and DFS
by Alan Biren 29 Apr '05

29 Apr '05

To add, NetApp can participate in a Windows DFS, but a key component is File Replication Service(FRS). NetApp does not particpate with the Windows Servers. If you are going to implement DFS and want the NetApp boxes to be in sync with your Windows Servers and eventually migrate off of the Windows File Servers VFM is the way to go. It'll manage the replication between Windows and DOT and share migration. Alan -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Charles Bartels Sent: Thursday, April 28, 2005 6:37 PM To: toasters(a)mathworks.com Subject: VFM and DFS Hi, My company currently has a couple of Netapps in the office along with a couple of Dell PC file servers. The windows admins would like to implement DFS. I don't really know what that is, but it sounded kind of like the Netapp VFM (Virtual File Manager). Could someone give me a short explanation of the two systems? Ideally I would like to consolidate our file servers onto the Netapps or at least avoid having to add unnecessary windows boxes. (These DFS servers sounded unnecessary to me, but if they are needed I wouldn't have a problem). Thanks. -C. ***************************************************************************************************** If you have received this email in error please notify the the system manager or the sender immediately and do not disclosed the to anyone or make copies. ** eSafe scanned this email for viruses, vandals or malicious content ** ** To find out eSafe(TM) and other Aladdin Security Products, contact AMC Corporation at (212) 736-0111 *****************************************************************************************************

1 0

SMB tuning/understanding
by Simon Waters 29 Apr '05

29 Apr '05

Pondering some Microsoft tuning guidelines in one window, and the Filer report in another. What is "OpLkBkDelayedBreak"? And better yet where do I find an idiots guide to the innards of Microsofts networking?

2 1

RE: VFM and DFS
by Palmer, Jason (EMEA) 29 Apr '05

29 Apr '05

Hi - VFM is a version of StorageX from Nuview (http://www.nuview.com which Network Appliance use to create a GSN (Global Storage Namespace) Nuview also has addition functionality that plugs into StorageX which is called ILM (Information Lifecycle Management) which takes advantage of virtulising the file system down to a file level. From what I can see there is an SQL database behing the application that holds information about where everything is in the storage environment. I have had StorageX installed on eval, and have recommended that we use it in favour of the Standard Windows DFS. It enables complex distributed storage environments to be presented, and manipulated very easily in comparison to the DFS with Windows. I hope this info is of help - Jason -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com]On Behalf Of Charles Bartels Sent: 28 April 2005 23:37 To: toasters(a)mathworks.com Subject: VFM and DFS Hi, My company currently has a couple of Netapps in the office along with a couple of Dell PC file servers. The windows admins would like to implement DFS. I don't really know what that is, but it sounded kind of like the Netapp VFM (Virtual File Manager). Could someone give me a short explanation of the two systems? Ideally I would like to consolidate our file servers onto the Netapps or at least avoid having to add unnecessary windows boxes. (These DFS servers sounded unnecessary to me, but if they are needed I wouldn't have a problem). Thanks. -C.

1 0

RE: Use minra or not?
by Derek Lai 29 Apr '05

29 Apr '05

Steve, Nice to have new toys to play with, isn't it? Minra is one of those things that depends heavily on your application's I/O pattern. If you have an application that is pretty sequential in I/O access, then you want read ahead. From experience, I can tell you that our Oracle Loan servicing application were beating the heck out of the disks we had. We were only using about 50% of the disk space but the utilization were very high. We turned on minra=on and right away cut about 30% disk utilization off. The good thing is that there were no noticeable cut in application performance. One other thing you might want to try (not sure how production it really is in your environment) is no_atime_update. Might help if your email application does not depend on it. Again, Oracle does not seem to need this. Derek -----Original Message----- From: Steve Losen [mailto:scl@sasha.acc.virginia.edu] Sent: Wednesday, April 27, 2005 3:03 PM To: toasters(a)mathworks.com Subject: Use minra or not? Please forgive all these questions to the list -- new toy syndrome. We have deployed a FAS960c pair in a University wide email system with about 30,000 accounts. We have 3 large Sun Solaris servers running Communigate Pro and they connect to the filer pair over a FC SAN. We have mapped 2 LUNs to each server, one from each filer, and striped them with Veritas. This has nicely balanced the load across both filers. We built a Veritas vxfs filesystem out of the two striped LUNs on each mail server and copied the inboxes over. Each mail server holds 1/3 of the inboxes and operates independently of the other two. It went into production today and the netapps appear to be more than up to the task. We are using mbox format where an email folder is one sequential file. Consequently our email servers read a LOT more KB/sec than they write, by a factor of more than ten to one. This afternoon I noticed that our filers were each reading about 100 MB/sec to 120 MB/sec from the disks. But I also noticed that this was 25% to 30% more than was being written to the FC SAN. Each filer has 6G of RAM. Reading disk data at a rate of 100MB/sec should completely turn over the RAM cache in about a minute, and sure enough, the cache age was typically 1 minute or less. I figured that we were reading 25% to 30% more from the disks than we were writing to the FC SAN because of read ahead. So I set "minra on" on the volume. Immediately we started reading a little less from the disks than we were writing to the FC SAN. Also disk util dropped from around 22% to around 15%. Cache hit percentage dropped a little from 97% to 96%, but then went back up. (How can cache hits be so high with the RAM cache being turned over every minute?) So it appears that in this situation that minra=on is a winner or is there something I have missed? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

RE: VFM and DFS
by Michael Schipp 29 Apr '05

29 Apr '05

The (possibly) best feature of DFS is that once created the underlining file shares can be moved from the current server(s) and consolidated on the Netapp (or any other file server) without having to change login script(s)/persistent connections etc. The DFS links just need to be changed. -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Sto Rage(c) Sent: Friday, 29 April 2005 10:22 AM To: Charles Bartels Cc: toasters(a)mathworks.com Subject: Re: VFM and DFS DFS (Distributed File System) - is a component of Microsoft Windows and runs on Windows 2000 or 2003 servers. Its a powerful feature but like everything else with Microsoft, they provide very basic tools to manage or implement it. Using DFS by itself was a disaster for us. VFM from Netapp - Provides an excellent wrapper around Microsoft's DFS, and provides lots of additional features. You can compare DFM - VFM to ntbackup - Netbackup from Veritas. Both DFS and ntbackup (builtin windows utilities) provides rudimentary features with hard to use GUI. Just enough to get you started. Whereas VFM and Netbackup provides a much better GUI and tons of other value adds. Hope you get the picture. We have been using DFS/VFM for almost 2 years and are very very happy wiith it. -G PS: Am just an end user and not affiliated to any of the vendors mentioned :) On 4/28/05, Charles Bartels <Charles.Bartels(a)tradebeam.com> wrote: > Hi, > > My company currently has a couple of Netapps in the office along with a > couple of Dell PC file servers. > > The windows admins would like to implement DFS. I don't really know > what that is, but it sounded kind of like the Netapp VFM (Virtual File > Manager). > > Could someone give me a short explanation of the two systems? Ideally I > would like to consolidate our file servers onto the Netapps or at least > avoid having to add unnecessary windows boxes. (These DFS servers > sounded unnecessary to me, but if they are needed I wouldn't have a > problem). > > Thanks. > > -C. >

1 0

VFM and DFS
by Charles Bartels 29 Apr '05

29 Apr '05

Hi, My company currently has a couple of Netapps in the office along with a couple of Dell PC file servers. The windows admins would like to implement DFS. I don't really know what that is, but it sounded kind of like the Netapp VFM (Virtual File Manager). Could someone give me a short explanation of the two systems? Ideally I would like to consolidate our file servers onto the Netapps or at least avoid having to add unnecessary windows boxes. (These DFS servers sounded unnecessary to me, but if they are needed I wouldn't have a problem). Thanks. -C.

2 1

Use minra or not?
by Steve Losen 27 Apr '05

27 Apr '05

Please forgive all these questions to the list -- new toy syndrome. We have deployed a FAS960c pair in a University wide email system with about 30,000 accounts. We have 3 large Sun Solaris servers running Communigate Pro and they connect to the filer pair over a FC SAN. We have mapped 2 LUNs to each server, one from each filer, and striped them with Veritas. This has nicely balanced the load across both filers. We built a Veritas vxfs filesystem out of the two striped LUNs on each mail server and copied the inboxes over. Each mail server holds 1/3 of the inboxes and operates independently of the other two. It went into production today and the netapps appear to be more than up to the task. We are using mbox format where an email folder is one sequential file. Consequently our email servers read a LOT more KB/sec than they write, by a factor of more than ten to one. This afternoon I noticed that our filers were each reading about 100 MB/sec to 120 MB/sec from the disks. But I also noticed that this was 25% to 30% more than was being written to the FC SAN. Each filer has 6G of RAM. Reading disk data at a rate of 100MB/sec should completely turn over the RAM cache in about a minute, and sure enough, the cache age was typically 1 minute or less. I figured that we were reading 25% to 30% more from the disks than we were writing to the FC SAN because of read ahead. So I set "minra on" on the volume. Immediately we started reading a little less from the disks than we were writing to the FC SAN. Also disk util dropped from around 22% to around 15%. Cache hit percentage dropped a little from 97% to 96%, but then went back up. (How can cache hits be so high with the RAM cache being turned over every minute?) So it appears that in this situation that minra=on is a winner or is there something I have missed? Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

RE: Aggregate snapshots?
by Fair, Robert 27 Apr '05

27 Apr '05

Aggregate snapshots are also used by the 'aggr copy' filer command, which is the aggregate-level version of 'vol copy'. This copies an aggregate and all its flexible volumes from one location to another, using aggregate snapshots to get a consistent image. -----Original Message----- From: Skottie Miller [mailto:skottie@anim.dreamworks.com] Sent: Wednesday, April 27, 2005 12:29 AM To: Derek Lai Cc: Steve Losen; toasters(a)mathworks.com Subject: Re: Aggregate snapshots? Derek Lai wrote: > Is this documented well anywhere? I've just started playing with > DataOntap 7.0 on the new FAS250 we bought. I've looked through NOW and > can't seem to find good documentation regarding Aggregate Snapshots. it's not documented well in any of the docs I have; my understanding is based on conversations with my SE and from a recent "7G migration" webcast. > It looks like there are a number of commands now that has -A options > added, i.e. snap list -A to list the aggregate snapshot. Would be nice > to know what Aggregate Snapshot are designed to do. There does not > seem to be anything in filerview that lets you work with Aggregate > snapshots? the use of -A is consistant throughput the command set; every command that you're using to using for a traditional volume works for volumes, and the -A version us ised to deal with the underlying aggregates. since we don't use RAID SyncMirror, my best practice for aggregate snapshots will be to disable them and to reclaim the space. If some compelling reason to use them can be explained to me, then I'll re-enable them. -skottie > > > Derek > > -----Original Message----- > From: Skottie Miller [mailto:skottie@anim.dreamworks.com] > Sent: Monday, April 25, 2005 11:53 AM > To: Steve Losen > Cc: toasters(a)mathworks.com > Subject: Re: Aggregate snapshots? > > > My understanding of aggregate snapshots is that they provide an > additional layer of snapshot protection, and are useful in some > environments. > > restoring an aggregate snapshot restores *all* the flexVols within > that aggregate. aggregate snapshots are required if you use RAID > SyncMirror and/or MetroCluster configurations. or of you want to > snaprestore a whole aggregate (seems dangerous). > > Note that aggregate snapshots aren't allowed to grow abouve the > reserve, and will "autodelete" themselves to maintain space. > > on my test filers and in my test use of the simulator, I've been > running with aggregate snapshots disabled and the reserve set to 0. > NetApp "discourages" this, but unless they send me 5% more disks, > free, I want that space back for production. > > -skottie > > Steve Losen wrote: > >>We just got a R200 and FAS960c and I immediately upgraded them from >>6.5 to 7.0.0.1. I moved the root volume to a flex vol and destroyed >>the old root, so now I have one big aggregate that contains the root >>volume. >> >>I discovered the "df -A" command and was surprised to see that my >>aggregate has snap reserve space, a snap sched running, and five >>snapshots. This is independent of the snap reserve and sched of my >>root volume. >> >>What is the rationale behind aggregate snapshots? We don't have >>snapmirror licensed. Can you snapmirror them? Can you snaprestore >>them? I presume this mirrors (or restores) the entire aggregate. >>I doubt that you can access aggregate snapshots from a NFS or CIFs >>client, or can you? Can you somehow tease out volumes from an >>aggregate snapshot and NFS export them? Within an aggregate snapshot, >>do you have just the active volumes, or the volumes and all of their >>volume level snapshots as they existed at the time the aggregate >>snapshot was taken? Mind boggling. >> >>I am probably going to set the snap reserve on my aggregates to zero >>and turn off the snap sched and manage my snapshots at the volume >>level. Any reason to not do that? Using both aggregate and volume >>snapshots looks about the same as using a belt and suspenders to keep >>your pants up. >> >>Steve Losen scl(a)virginia.edu phone: 434-924-0640 >> >>University of Virginia ITC Unix Support >> >> > >

1 0

list admin page
by Robert Borowicz 27 Apr '05

27 Apr '05

Can someone forward me the "list admin page" for the Toasters mailing list. I need to move my subscription to my new ISP. Thanks -Rob

1 0

RE: Aggregate snapshots?
by Ehrhart, Rick 27 Apr '05

27 Apr '05

FilerView for ONTAP 7.0 concentrated on the commonly used features of Flexible Volumes and Aggregrates. Aggregate snapshots was not considered a commonly used feature. - Rick - > -----Original Message----- > From: Derek Lai [mailto:Derek.Lai@onyxco.com] > Sent: Tuesday, April 26, 2005 6:12 PM > To: 'Skottie Miller'; Steve Losen > Cc: toasters(a)mathworks.com > Subject: RE: Aggregate snapshots? > > > Is this documented well anywhere? I've just started playing > with DataOntap > 7.0 on the new FAS250 we bought. I've looked through NOW and > can't seem to > find good documentation regarding Aggregate Snapshots. > > It looks like there are a number of commands now that has -A > options added, > i.e. snap list -A to list the aggregate snapshot. Would be > nice to know what > Aggregate Snapshot are designed to do. There does not seem to > be anything in > filerview that lets you work with Aggregate snapshots? > > > > Derek > > -----Original Message----- > From: Skottie Miller [mailto:skottie@anim.dreamworks.com] > Sent: Monday, April 25, 2005 11:53 AM > To: Steve Losen > Cc: toasters(a)mathworks.com > Subject: Re: Aggregate snapshots? > > > My understanding of aggregate snapshots is that they provide > an additional > layer of snapshot protection, and are useful in some environments. > > restoring an aggregate snapshot restores *all* the flexVols > within that > aggregate. aggregate snapshots are required if you use RAID > SyncMirror > and/or MetroCluster configurations. > or of you want to snaprestore a whole aggregate (seems dangerous). > > Note that aggregate snapshots aren't allowed to grow abouve > the reserve, and > will "autodelete" themselves to maintain space. > > on my test filers and in my test use of the simulator, I've > been running > with aggregate snapshots disabled and the reserve set to 0. NetApp > "discourages" this, but unless they send me 5% more disks, > free, I want that > space back for production. > > -skottie > > Steve Losen wrote: > > We just got a R200 and FAS960c and I immediately upgraded them from > > 6.5 to 7.0.0.1. I moved the root volume to a flex vol and > destroyed > > the old root, so now I have one big aggregate that contains > the root > > volume. > > > > I discovered the "df -A" command and was surprised to see that my > > aggregate has snap reserve space, a snap sched running, and five > > snapshots. This is independent of the snap reserve and sched of my > > root volume. > > > > What is the rationale behind aggregate snapshots? We don't have > > snapmirror licensed. Can you snapmirror them? Can you snaprestore > > them? I presume this mirrors (or restores) the entire aggregate. > > I doubt that you can access aggregate snapshots from a NFS or CIFs > > client, or can you? Can you somehow tease out volumes from an > > aggregate snapshot and NFS export them? Within an > aggregate snapshot, > > do you have just the active volumes, or the volumes and all > of their > > volume level snapshots as they existed at the time the aggregate > > snapshot was taken? Mind boggling. > > > > I am probably going to set the snap reserve on my > aggregates to zero > > and turn off the snap sched and manage my snapshots at the volume > > level. Any reason to not do that? Using both aggregate and volume > > snapshots looks about the same as using a belt and > suspenders to keep > > your pants up. > > > > Steve Losen scl(a)virginia.edu phone: 434-924-0640 > > > > University of Virginia ITC Unix Support > > > > > > -- > > Scott Miller > skottie(a)DreamWorksAnimation.com >

1 0

RE: Aggregate snapshots?
by Derek Lai 27 Apr '05

27 Apr '05

Is this documented well anywhere? I've just started playing with DataOntap 7.0 on the new FAS250 we bought. I've looked through NOW and can't seem to find good documentation regarding Aggregate Snapshots. It looks like there are a number of commands now that has -A options added, i.e. snap list -A to list the aggregate snapshot. Would be nice to know what Aggregate Snapshot are designed to do. There does not seem to be anything in filerview that lets you work with Aggregate snapshots? Derek -----Original Message----- From: Skottie Miller [mailto:skottie@anim.dreamworks.com] Sent: Monday, April 25, 2005 11:53 AM To: Steve Losen Cc: toasters(a)mathworks.com Subject: Re: Aggregate snapshots? My understanding of aggregate snapshots is that they provide an additional layer of snapshot protection, and are useful in some environments. restoring an aggregate snapshot restores *all* the flexVols within that aggregate. aggregate snapshots are required if you use RAID SyncMirror and/or MetroCluster configurations. or of you want to snaprestore a whole aggregate (seems dangerous). Note that aggregate snapshots aren't allowed to grow abouve the reserve, and will "autodelete" themselves to maintain space. on my test filers and in my test use of the simulator, I've been running with aggregate snapshots disabled and the reserve set to 0. NetApp "discourages" this, but unless they send me 5% more disks, free, I want that space back for production. -skottie Steve Losen wrote: > We just got a R200 and FAS960c and I immediately upgraded them from > 6.5 to 7.0.0.1. I moved the root volume to a flex vol and destroyed > the old root, so now I have one big aggregate that contains the root > volume. > > I discovered the "df -A" command and was surprised to see that my > aggregate has snap reserve space, a snap sched running, and five > snapshots. This is independent of the snap reserve and sched of my > root volume. > > What is the rationale behind aggregate snapshots? We don't have > snapmirror licensed. Can you snapmirror them? Can you snaprestore > them? I presume this mirrors (or restores) the entire aggregate. > I doubt that you can access aggregate snapshots from a NFS or CIFs > client, or can you? Can you somehow tease out volumes from an > aggregate snapshot and NFS export them? Within an aggregate snapshot, > do you have just the active volumes, or the volumes and all of their > volume level snapshots as they existed at the time the aggregate > snapshot was taken? Mind boggling. > > I am probably going to set the snap reserve on my aggregates to zero > and turn off the snap sched and manage my snapshots at the volume > level. Any reason to not do that? Using both aggregate and volume > snapshots looks about the same as using a belt and suspenders to keep > your pants up. > > Steve Losen scl(a)virginia.edu phone: 434-924-0640 > > University of Virginia ITC Unix Support > > -- Scott Miller skottie(a)DreamWorksAnimation.com

2 1

RE: How does FCP cfmode partner work?
by Hill, Aaron 27 Apr '05

27 Apr '05

Yeah you have pretty much worked it out. For FCP Partner cfmode, the filer uses a feature called "FCP Proxy" to forward requests to the partner LUN. As you said, in normal mode, primary paths are through the A port on Filer 1 and B'' port (second adapter) of Filer2. The FCP Proxy allows MPIO s/w to maintain active-passive connection by using the VTIC adapter between the two filers to proxy commands. As far as I know, your host requires MPIO software to achieve this. I don't think this has changed for 7G, but I haven't reviewed it yet. Thie whole thing is possible because the B ports WWPN/WWNN are based on the Partner nodes WWNN. Thus for AIX/HP-UX types, the requirement that static S_ID and D_ID are maintained because the WWPN/WWNN don't change during failover. This keeps the MPIO software happy. No changes on the switch are made. Both access the Filer1 LUNs via the A port on Filer1 and B port on Filer2 (FCP Proxy does the forwarding) B port WWNN/WWPN based on partner WWPN. Hope that explains it. Now, if you start doing mixed cfmode using the public loop/multi-ID, it will take me a little longer to explain. --Aaron -----Original Message----- From: Steve Losen [mailto:scl@sasha.acc.virginia.edu] Sent: Wednesday, 27 April 2005 8:23 AM To: toasters(a)mathworks.com Subject: How does FCP cfmode partner work? We recently set up a new FAS960c pair of filers running 7.0.0.1 and I am trying to understand how cfmode partner works. According to the "Block Access Management Guide for FCP" the A port on each FC adapter is for accessing local LUNs while the B port is for accessing partner LUNs. In "cfmode partner" both ports are active. So does that mean that in normal operation (no failover) that initiators (hosts) can access a LUN on filer 1 from a B port on filer 2? That is what the manual seems to indicate. How is this accomplished? Does it happen within the switch? Do FCP requests for a B port get forwarded to an A port on the partner during normal operation? Obviously during a failover the surviving head takes over the disks of its partner and then satisfies requests to the B port itself. Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support ************** IMPORTANT MESSAGE ************** This e-mail message is intended only for the addressee(s) and contains information which may be confidential. If you are not the intended recipient please advise the sender by return email, do not use or disclose the contents, and delete the message and any attachments from your system. Unless specifically indicated, this email does not constitute formal advice or commitment by the sender or the Commonwealth Bank of Australia (ABN 48 123 123 124) or its subsidiaries. We can be contacted through our web site: commbank.com.au. If you no longer wish to receive commercial electronic messages from us, please reply to this e-mail by typing Unsubscribe in the subject line. ***************************************************************

1 0

How does FCP cfmode partner work?
by Steve Losen 26 Apr '05

26 Apr '05

We recently set up a new FAS960c pair of filers running 7.0.0.1 and I am trying to understand how cfmode partner works. According to the "Block Access Management Guide for FCP" the A port on each FC adapter is for accessing local LUNs while the B port is for accessing partner LUNs. In "cfmode partner" both ports are active. So does that mean that in normal operation (no failover) that initiators (hosts) can access a LUN on filer 1 from a B port on filer 2? That is what the manual seems to indicate. How is this accomplished? Does it happen within the switch? Do FCP requests for a B port get forwarded to an A port on the partner during normal operation? Obviously during a failover the surviving head takes over the disks of its partner and then satisfies requests to the B port itself. Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

command usage summary
by Tim Thompson 26 Apr '05

26 Apr '05

To inform decisions about future products, we at Network Appliance would like to get a feel for current command usage by customers, particularly as used from scripts. Unfortunately, autosupport doesn't contain the level of detail that we'd like to see. Attached below is a perl script that will look at the auditlogs of a filer running Data ONTAP and summarize command, sub-command, and api usage. Only a summary is produced - it shouldn't contain any customer-sensitive information. It will not put any load on the filer, and should consume only a minute or so of CPU time on the machine running the perl script (depending on how large your auditlog files are). Here's a small sample of the output: count=24 rsh=vol options count=9 rsh=vol status count=6 rsh=version count=5 rsh=sysconfig -a count=3 rsh=vol create count=3 rsh=reboot -t count=11 cons=version count=6 cons=options ldap count=22 api=perf-object-instance-list-info count=16 api=perf-object-counter-list-info count=13 api=options-get We would greatly appreciate it if anyone using Data ONTAP could run this script and send the output to usage(a)netapp.com (a special address set up to collect this data). Usage is described in the comments at the top of the file. Thanks much. ...Tim Thompson...Network Appliance, Inc...tjt(a)netapp.com... #----------------------------- cut here ----------------------------- #!/usr/local/bin/perl # # This script summarizes command and api usage on a # Network Appliance filer running Data ONTAP. In other words, it counts # how many times each command/sub-command has been used, by # looking at the auditlogs. # # Usage on Unix is: # # cd /mnt/$FILER/etc # where /mnt/$FILER is the mounted root volume # perl cmdsummary.pl # # or on Windows: # # cd F: # where F: is the mounted root volume # cd \etc # perl cmdsummary.pl # # In other words, change directories to the /etc directory of the # filer's root volume, and run this script. The script looks at # the messages file and all the auditlog files, and summarizes # command and api usage. my %rshusage; my %consusage; my %apiusage; my $system = ""; my $anyerr = 0; foreach $f (<messages log/auditlog*>) { if ( ! open(FF, $f) ) { print STDERR "Error: unable to open $f - $!\n\n" ."This program must be run in the " ."'etc' directory of the filer's root volume.\n\n"; $anyerr = 1; next } while (<FF>) { if ( /kern.log.rotate:/ || /log_rotate:/ ) { chop; $system = $_; $system =~ s/.*System /System /; } elsif ( /RSH INPUT COMMAND/ ) { my $s = $_; $s =~ s/.*RSH INPUT COMMAND is //; my @words = split(/[\s;]/,$s); my $cmd = $words[0]." ".$words[1]; if ( $words[0] ne "" ) { if ( not defined $rshusage{$cmd} ) { $rshusage{$cmd} = 1; } else { $rshusage{$cmd} += 1; } } } elsif ( /IN:console shell:/ ) { my $s = $_; $s =~ s/.*console shell://; my @words = split(/\s/,$s); my $cmd = $words[0]." ".$words[1]; if ( $words[0] ne "" ) { if ( not defined $consusage{$cmd} ) { $consusage{$cmd} = 1; } else { $consusage{$cmd} += 1; } } } elsif ( /Java_Thread.*:API:in:/ && ! /relog syslog/ ) { chop; my $api = $_; $api =~ s/[^<]*<//; $api =~ s/\/?>.*//; if ( not defined $apiusage{"Java-$api"} ) { $apiusage{"Java-$api"} = 1; } else { $apiusage{"Java-$api"} += 1; } } elsif ( /:API:in:/ && ! /relog syslog/ ) { chop; my $api = $_; $api =~ s/.*\/admin\'><//; $api =~ s/.*version="..."><//; $api =~ s/\/?>.*//; if ( not defined $apiusage{$api} ) { $apiusage{$api} = 1; } else { $apiusage{$api} += 1; } } } } if ( $system ne "" ) { print("$system\n"); } foreach $key ( sort { $rshusage{$b} <=> $rshusage{$a} } keys %rshusage ) { print("count=$rshusage{$key} rsh=$key\n"); } foreach $key ( sort { $consusage{$b} <=> $consusage{$a} } keys %consusage ) { print("count=$consusage{$key} cons=$key\n"); } foreach $key ( sort { $apiusage{$b} <=> $apiusage{$a} } keys %apiusage ) { print("count=$apiusage{$key} api=$key\n"); } exit($anyerr); #----------------------------- end of script -----------------------------

1 0

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