toasters

toasters@lists.teaparty.net

13531 discussions

RE: Cifs.max_mpx
by Hawley, Rob 17 Apr '00

17 Apr '00

Cifs.max_mpx tells the client how many operations can be pending on the filer at the same time. The value should never be set below the NT default of 50 (some early versions of 5.3 set it to 32 which was incorrect). If cifs stat shows Max Multiplex greater than about 32 you need to raise this number. The two engineering approved increases are 253 and 1124. Do not set this to any other values without an explicit conversation with NetApp customer service. One way you can tell how much is enough is to run perfmon on the client consuming the most resources and look at the Redirector\current commands stat. Max_mpx should be set to a healthy margin above that number. It should be noted that just automatically setting this high is not a good idea. High values consume a lot of resources in the clients. That is one of the reasons this is a hidden option. Setting this to an unapproved value will cause errors on some clients http://support.microsoft.com/support/kb/articles/Q232/8/90.ASP. Remember that options.cifs.max_mpx controls a number passed in the initial connection of a client to the filer. Changing this value will not affect active sessions. For this reason we recommend only changing this option when cifs is terminated. Once cifs restarts you can insure that all connections will have the new value There is a similar parameter on an NT server. See http://support.microsoft.com/support/kb/articles/q191/3/70.asp -----Original Message----- From: ZANGA, Michael, GCM [mailto:Michael.Zanga@gcm.com] Sent: Monday, April 17, 2000 8:16 AM To: toasters(a)mathworks.com Subject: Cifs.max_mpx Does anyone have any idea what the option cifs.max_mpx does. On a whitepaper from NetApp the option is referenced to be changed to 1124. I presently have it set at 32 and it was hidden. I am reluctant to make the change without knowing what it does. ********************************************************************** This e-mail is intended only for the addressee named above. As this e-mail may contain confidential or privileged information, if you are not the named addressee, you are not authorised to retain, read, copy or disseminate this message or any part of it. ************************************************************************

1 0

RE: Filer volume configuration question
by Norman, Paul 17 Apr '00

17 Apr '00

Steve said "one volume per head", which means the Clustered pair must always have a minimum of two volumes. With a single volume per filer (root volume), each filer is active and serves data from its designated volume. If there is a takeover, then the remaining filer serves data from both volumes. You can have a "hot spare filer", by creating a small root volume on one filer and not have that filer serve any data, but as you say, two volumes each of which has data to be served will give you the best bang for the buck. Paul Norman, Product Manager Direct: (408) 822-6288 Network Appliance E-Mail: pnorman(a)netapp.com 495 East Java Drive Corp: (408) 822-6000 Sunnyvale, CA 94089 Web: WWW.NetApp.Com > -----Original Message----- > From: July at Zerowait [mailto:july@zerowait.com] > Sent: Monday, April 17, 2000 8:05 AM > To: Watanabe, Steve > Cc: 'ari(a)genuity.net'; toasters(a)mathworks.com > Subject: Re: Filer volume configuration question > > > Now I am confused. If there is only one volume, then what you have is > essentially a "hot spare filer" for failover. Surely setting > up two volumes > lets you use both filers actively and gets you more bang for > the buck? While > a single volume cluster might work, is it really the > recommended config? > > July > > "Watanabe, Steve" wrote: > > > Hi Ari, > > A cluster requires a minimum of one volume per head. So > having a 6 disk > > vol0 > > and a spare per node is fine. Make sure that you follow > the directions in > > the > > cluster configuration guide carefully to avoid any > situation where both > > nodes > > are cabled incorrectly allowing access to the same volume > at the same > > time. In fact, I thought NetApp always did the initial cluster > > configuration. > > Is that option available to you? > > -Steve > > > > -----Original Message----- > > From: Ari [mailto:ari@genuity.net] > > Sent: Monday, April 17, 2000 5:31 AM > > To: toasters(a)mathworks.com > > Subject: Filer volume configuration question > > > > Hi, > > > > I am setting up a pair of clustered 740's for the first > time and have a > > question. > > > > Each filer currently has one disk shelf. 6 of the disks > are in the root > > volume, /vol/vol0, and the 7th is a hot spare. Is this > configuration OK > > for a clustered pair? I was told by someone that for a > cluster, I would > > need 2 distinct volumes on each head, although I am not > entirely sure why. > > If that is the case, how do I remove drives from the root > volume so that I > > can add them into a newly created volume? > > > > Any help is greatly appreciated. > > > > > > Ari > > > > > > > > > > Ari Globerman > > > > eBusiness Web Hosting Services > > e-mail: ari(a)genuity.com <mailto:ari@genuity.com> > > phone: 617-873-5843 > > fax: 617-873-4033 > > > > <http://www.genuity.com/> > > > > > > > > > -------------------------------------------------------------- > ---------- > > Name: GTEI.h1.gif > > GTEI.h1.gif Type: GIF Image (image/gif) > > Encoding: base64 > > -- > ---------------------------------------------------------------- > July A. J. Linett, VP Tech Mktg > 302.266.9408 july(a)zerowait.com > > Zerowait: Best Of Breed Internet Core Technologies > Web Acceleration ** Data Storage ** Load Balancing > www.zerowait.com*www.loadbalancing.net*www.nas-san.com > ------------------------------------------------------------------- > >

1 0

RE: Filer volume configuration question
by Watanabe, Steve 17 Apr '00

17 Apr '00

Hi Ari, A cluster requires a minimum of one volume per head. So having a 6 disk vol0 and a spare per node is fine. Make sure that you follow the directions in the cluster configuration guide carefully to avoid any situation where both nodes are cabled incorrectly allowing access to the same volume at the same time. In fact, I thought NetApp always did the initial cluster configuration. Is that option available to you? -Steve -----Original Message----- From: Ari [mailto:ari@genuity.net] Sent: Monday, April 17, 2000 5:31 AM To: toasters(a)mathworks.com Subject: Filer volume configuration question Hi, I am setting up a pair of clustered 740's for the first time and have a question. Each filer currently has one disk shelf. 6 of the disks are in the root volume, /vol/vol0, and the 7th is a hot spare. Is this configuration OK for a clustered pair? I was told by someone that for a cluster, I would need 2 distinct volumes on each head, although I am not entirely sure why. If that is the case, how do I remove drives from the root volume so that I can add them into a newly created volume? Any help is greatly appreciated. Ari Ari Globerman eBusiness Web Hosting Services e-mail: ari(a)genuity.com <mailto:ari@genuity.com> phone: 617-873-5843 fax: 617-873-4033 <http://www.genuity.com/>

2 1

RE: NetApp F7XX + Snapmirror
by Walters, Mike 17 Apr '00

17 Apr '00

Just an additional thought: You can of course mirror vol0 (your 'root + data' volume) from Filer A to vol1 on Filer B as Stephen states. You can then, if you wish, mirror the reverse as well - in other words, have another vol1 on Filer A, & mirror Filer B's 'root + data' to it. I know you never have enough to read ;-) so for an overview of the SnapMirror (& SnapRestore) functionality, you might want to take a look at http://www.netapp.com/tech_library/3043.html. The NOW site has a more in-depth operational guide at http://now.netapp.com/NOW/knowledge/docs/olio/guides/TR3073.shtml. (& as always, the SAG is a good reference point for how-to information ;-) Cheers Mike > First, ALL of a snapmirrored volume is mirrored. There is no way > to make it finer-grained, or to make it exclude certain files or > directories. > > Let's say you have machines A and B. > > A has one volume: vol0. > vol0 has 7 disks and is the root volume. Its /etc files are > used for machine configuration. > > B has two volumes: vol0 and vol1. > vol0 has 2 disks and is the root volume. Its /etc files are > used for machine configuration. > vol1 has 7 disks and is a snapmirror of A:vol0. It is > only readable, and its /etc files are NOT used for > machine configuration, unless you break the > mirror, make it > writeable, and make it the root volume. > > > So, I don't think you need to be too worried about the mirroring of > /etc. However, you still would need 2 more disks on B:vol1 > to be able > to mirror A:vol0 to it. > > Stephen Manley > Data Availability and Management Set-Up Reliever >

2 1

RE: Filer volume configuration question
by Sims, Aaron 17 Apr '00

17 Apr '00

The configuration you specified is perfectly fine for a cluster install. You do not need two volumes for CFO. You do need two volumes for a destination filer in a SnapMirror configuration. Aaron -- Aaron Sims Professional Services Engineer Network Appliance aaron.sims(a)netapp.com 404-226-7382 (cell) 408-822-8947 (voice mail) -----Original Message----- From: Ari [mailto:ari@genuity.net] Sent: Monday, April 17, 2000 8:31 AM To: toasters(a)mathworks.com Subject: Filer volume configuration question Hi, I am setting up a pair of clustered 740's for the first time and have a question. Each filer currently has one disk shelf. 6 of the disks are in the root volume, /vol/vol0, and the 7th is a hot spare. Is this configuration OK for a clustered pair? I was told by someone that for a cluster, I would need 2 distinct volumes on each head, although I am not entirely sure why. If that is the case, how do I remove drives from the root volume so that I can add them into a newly created volume? Any help is greatly appreciated. Ari Ari Globerman eBusiness Web Hosting Services e-mail: ari(a)genuity.com <mailto:ari@genuity.com> phone: 617-873-5843 fax: 617-873-4033 <http://www.genuity.com/>

1 0

RE: Filer volume configuration question
by Cody, Jason 17 Apr '00

17 Apr '00

Ari- You do NOT need more than one volume on each filer to install a cluster. Your configuration is fine. The person you were talking to may have been thinking about snapmirror. That requires more than one volume to be successful. hope this helps- Jason Cody Network Appliance Professional Services Engineer East Mid-Atlantic Region jcody(a)netapp.com 410-299-6290 Cell 888-374-7504 Pager 408-822-8895 Voice Mail 801-340-7018 eFax -----Original Message----- From: Ari [mailto:ari@genuity.net] Sent: Monday, April 17, 2000 8:31 AM To: toasters(a)mathworks.com Subject: Filer volume configuration question Hi, I am setting up a pair of clustered 740's for the first time and have a question. Each filer currently has one disk shelf. 6 of the disks are in the root volume, /vol/vol0, and the 7th is a hot spare. Is this configuration OK for a clustered pair? I was told by someone that for a cluster, I would need 2 distinct volumes on each head, although I am not entirely sure why. If that is the case, how do I remove drives from the root volume so that I can add them into a newly created volume? Any help is greatly appreciated. Ari Ari Globerman eBusiness Web Hosting Services e-mail: ari(a)genuity.com <mailto:ari@genuity.com> phone: 617-873-5843 fax: 617-873-4033 <http://www.genuity.com/>

1 0

Filer volume configuration question
by Ari 17 Apr '00

17 Apr '00

Hi, I am setting up a pair of clustered 740's for the first time and have a question. Each filer currently has one disk shelf. 6 of the disks are in the root volume, /vol/vol0, and the 7th is a hot spare. Is this configuration OK for a clustered pair? I was told by someone that for a cluster, I would need 2 distinct volumes on each head, although I am not entirely sure why. If that is the case, how do I remove drives from the root volume so that I can add them into a newly created volume? Any help is greatly appreciated. Ari Ari Globerman eBusiness Web Hosting Services e-mail: ari(a)genuity.com phone: 617-873-5843 fax: 617-873-4033

1 0

large disk activity but no net io or nfs?
by mgx＠ornl.gov 17 Apr '00

17 Apr '00

hmmm ... just having a look at sysstat on my F520's this morning: CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache in out read write read write age 79% 1 0 0 2 0 11619 0 0 0 >60 79% 0 0 0 2 0 11314 52 0 0 >60 75% 1 0 0 2 0 11158 0 0 0 >60 76% 8 0 0 48 2 11517 0 0 0 >60 83% 17 0 0 6 3 11700 157 0 0 >60 76% 2 0 0 3 0 11199 0 0 0 >60 80% 12 0 0 4 15 11743 63 0 0 >60 74% 1 0 0 2 0 11329 0 0 0 >60 74% 1 0 0 3 0 10791 45 0 0 >60 72% 2 0 0 3 0 10904 1 0 0 >60 73% 1 0 0 2 0 11728 0 0 0 >60 seems like a lot of CPU and DISK IO for no nfs ops. any idea what could be going on here? thanks! -- michael

4 5

NetApp F7XX + Snapmirror
by TTSG 16 Apr '00

16 Apr '00

Hi, Thanks to all those who helped me over the Networking problem... Now on to bigger/better/harder things. As I mentioned, the customer had a F720 shipped in. When I configured it, I just threw everything into vol0. Now when I look to use SNAPMIRROR to mirror from "A" filer to "B" filer, I can do it by volumes only. Am I correct isn ASSuM(E)ing that if I mirror "vol0" I'll clobber the configuration of "B" filer? I decided to believe this was correct, and reloaded the "B" filer. I created a "vol1" on it with the remaining 5 disks. In attempting to setup the snapmirror, it now complains that the volume is on-line. Um, how is it going to let people still access it if it goes offline? So to see where I can get myself into more trouble, I offline it, reboot, and now its complaining they are different sizes. I have a feeling there is A BUNCH of confusion on my part. Maybe if I say what I want, someone can shoot the clue-gun at me. I want only 1 computer to have update access to filer "A". This will be the host that updates the content on the disk. 3 other computers will have this mounted RO. I will also have filer "B". 4 computers will have this mounted RO. I'd like to have "A" filer copy the content to "B" filer once every (??? not sure yet, say minute to be on the safe side ???). Thanks, Tuc/TTSG

2 3

1400 model
by John Stacey 16 Apr '00

16 Apr '00

IM sure this has come up before, but has anyone attempted overclocking any of the older model filers? Just curious how well they can (or cant) take the upgrades commonly available for 486 based systems. Thanks

3 2

Problem in Automounting
by bhaskar.g＠philips.com 15 Apr '00

15 Apr '00

Hi all, Here is a real good problem. I shall explain it to you in detail. We have a filer F720 with 2*18GB*7disks shelves and we have configure a single volume on that /vol/vol0 We have raid 4 configured on it. The total capacity of the vol0 is about 150GB. We have started dumping data on it and almost crossed 70GB. We have 3 NIS indirect maps for mounting of the directories one for the user home directories one for Project directories one for System Binary directories. All the mount points were accessible till yesterday. Problem is as follows: we had to change the subnet mask of the whole network from 255.255.255.0 to 255.255.254.0 We have done change of netmaks on all the active switches, servers, clients and also the filer. The problem started now Some of the users home directories were not accessible from the this maps but for me and some others there is no problem at all accessing the data root@austin# cd /home/www root@austin# ll . not found root@austin# bdf . bdf: .: Permission denied root@austin# root@austin# cd /home/bhaskarg root@austin# ls .ICAClient NetHood fmfilesvisited .ICEauthority PC Database.xls frame .WindU Personal friend.doc .Xauthority Pop3Server-2_14.zip friendship.pps .acrorc PrintHood gcc-2.95.2-sol7-sparc-local .acrosrch RECEIPT.doc gcc-2.95.2.tar .addrbk Recent general.asp .addressbook SOFT_OF2.DOC ghostscript-5.50-sol7-sparc-local root@austin# root@austin# bdf . Filesystem kbytes used avail %used Mounted on drona:/vol/vol0/home/bhaskarg 150223752 72078752 78145000 48% /tmp_mnt/home/bhaskarg root@austin# when I tried to change the user www's home directory to some other unix system I am able to see the files in the directory. This means that there is not problem in the NIS Indirect map file. Do anyone have solution for this problem or anyone has come across such a type of problem. Please help us in solving the problem Thanks & Regards Bhaskar G Systems Engineer, Philips Semiconductors, #3, Sterling Square,Madras Bank Road, Off St Mark's Road,BANGALORE-560 001. Voice : +91 80 222 9200 Ext 1478 Fax: +91 80 2229700

1 0

RE: IIS and netapp in existing unix world
by Benn, Paul 15 Apr '00

15 Apr '00

Clarification: With the later versions of Data ONTAP you can leave oplocks on. Regards, Paul -----Original Message----- From: Benn, Paul [mailto:paul.benn@netapp.com] Sent: Friday, April 14, 2000 2:11 PM To: 'Mathew Anderson'; toasters(a)mathworks.com Subject: RE: IIS and netapp in existing unix world There are a number of NetApp customers back ending their IIS servers with filers. Make sure you disable oplocks if you are pointing more than one IIS machine to the same data on the filer. When using IIS 4.0, the recommended release of Data ONTAP is 5.3.5P2 or later. You can configure the home directory with any user account you wish; in the lab we usually use the account that the IIS installation app creates. Just make sure it has the appropriate share level and NTFS level permissions (usually read-only for IIS). As I recall, the IUSER* account is configured without a password (I like to change that) and IIS will let you configure the home directory with an incorrect password (even though it may match when you're prompted to verify it). You should be fine using a qtree with mixed permissions. We just published an IIS white paper which should be helpful for initial setup. It can be found at: http://www.netapp.com/tech_library/3078.html Regards, Paul Benn Network Appliance -----Original Message----- From: Mathew Anderson [mailto:mathew@nm.net] Sent: Friday, April 14, 2000 1:35 PM To: toasters(a)mathworks.com Subject: IIS and netapp in existing unix world Hello -- I have a 740 in production in a unix environment. I am looking forward to using this as a backend server for my NT web farm. I created a qtree and set it with mixed permissions (I also would like to set my linux web farm up and mount the same partitions). The problem that I am having is that IIS 4, (on NT4-sp6a) gives the error: Unable to enumerate files and directories because the following error occurred: Access is denied for the home directory setting, I am using a network directory and connecting to the share. I can do the a net view \\machine without a problem. An additional question is, what user should I be connecting as? and with unix permissions set for the volume (and mixed for the qtree) should there be a problem with IIS looking at the mount? I can mount the filesystem without a problem. Thanks for any help. Mathew Anderson Systems Administration NM Technet Oso Grande Technologies, Inc. mathew(a)nm.net mathew(a)osogrande.net

1 0

RE: How do I move /vol/vol0?
by Brookes, Matt 15 Apr '00

15 Apr '00

Actually, you don't have to destroy the old volume until you're happy with the transition, though you might want to offline it. It's also possible to rename the old volume to something other than vol0, and to rename the new volume to vol0, this should save some hassle with exports & other config info. You may still run in to problems with stale file handles if you have clients with active mounts, but I guess since this isn't something you'll be doing regularly that shouldn't be a big deal. (That's why management created weekends, right? ;-) Regards, Matt. ________________________________________________________________ Matthew Brookes mailto:mbrookes@netapp.com Systems Engineer http://www.netapp.com Network Appliance (Sales) Ltd. Tel: +353 1 4064600 Rathfarnham Gate Fax: +353 1 4064601 Dublin 14, IRELAND Mob: +353 86 8575127 "Anyone looking for storage on the Net turns to Network Appliance" Fortune, December 1999. > -----Original Message----- > From: Fox, Adam [mailto:Adam.Fox@netapp.com] > Sent: 14 April 2000 19:42 > To: 'Cliff Nadler'; toasters(a)mathworks.com > Subject: RE: How do I move /vol/vol0? > > > > > > I'm in a situatio where we have one F760 with 4 shelves of > > 9Gb SCSI disk, and > > three shelves of 36Gb FC attached disks. The FC disks were > > added later, so the > > original install of vol0 was done to the SCSI disks. > > > > We'd like to get rid of the SCSI disks (and use one shelf of > > FC to hold it), > > but I'm concerned since we have /vol/vol0 on it. Has anyone > > moved /vol/vol0 off > > SCSI to FC before, and if so, what problems did you run into? > > > > Many customers have used vol copy to move the data over. Just > make sure that the destination fiber-based volume is at least as > large as the original SCSI volume. You vol copy the data over. > Set the new volume to be the root volume. Then you can destroy > the SCSI volume. I'd probably type download, just to be safe. > Then I'd reboot...the filer won't actually destroy the old vol0 > and start using the new volume until you actually reboot. Then > if you want to keep the vol0 name, you can rename the new root > volume. > > This has worked for many customers that I've talked to. Has anyone > seen differently? > > -- Adam Fox > NetApp Professional Services, NC > adamfox(a)netapp.com >

1 0

RE: Disk-to-disk backups? (was Re: Need new backup solution)
by mitch＠netline.com 14 Apr '00

14 Apr '00

/* Andrew Bond [abond(a)netapp.com] writes: */ >I'm not clear how a hacker could destroy data in snapshots - remember that >they are read-only, and that's why they have proved useful in cases of >virus attack. > I can't say what Bruce was thinking, but a cracker could do the obvious: snap delete <vol-name> name once they gain access to the filer either directly or via the admin host. That pretty much destroys data in a snapshot. :-) -- do svidaniya, ~mitch

1 0

RE: IIS and netapp in existing unix world
by Benn, Paul 14 Apr '00

14 Apr '00

There are a number of NetApp customers back ending their IIS servers with filers. Make sure you disable oplocks if you are pointing more than one IIS machine to the same data on the filer. When using IIS 4.0, the recommended release of Data ONTAP is 5.3.5P2 or later. You can configure the home directory with any user account you wish; in the lab we usually use the account that the IIS installation app creates. Just make sure it has the appropriate share level and NTFS level permissions (usually read-only for IIS). As I recall, the IUSER* account is configured without a password (I like to change that) and IIS will let you configure the home directory with an incorrect password (even though it may match when you're prompted to verify it). You should be fine using a qtree with mixed permissions. We just published an IIS white paper which should be helpful for initial setup. It can be found at: http://www.netapp.com/tech_library/3078.html Regards, Paul Benn Network Appliance -----Original Message----- From: Mathew Anderson [mailto:mathew@nm.net] Sent: Friday, April 14, 2000 1:35 PM To: toasters(a)mathworks.com Subject: IIS and netapp in existing unix world Hello -- I have a 740 in production in a unix environment. I am looking forward to using this as a backend server for my NT web farm. I created a qtree and set it with mixed permissions (I also would like to set my linux web farm up and mount the same partitions). The problem that I am having is that IIS 4, (on NT4-sp6a) gives the error: Unable to enumerate files and directories because the following error occurred: Access is denied for the home directory setting, I am using a network directory and connecting to the share. I can do the a net view \\machine without a problem. An additional question is, what user should I be connecting as? and with unix permissions set for the volume (and mixed for the qtree) should there be a problem with IIS looking at the mount? I can mount the filesystem without a problem. Thanks for any help. Mathew Anderson Systems Administration NM Technet Oso Grande Technologies, Inc. mathew(a)nm.net mathew(a)osogrande.net

1 0

Ontap 5.4
by Jay Orr 14 Apr '00

14 Apr '00

Any word if ontap 5.4 will be out this month? The reason I ask is Netapp told us they are discontinuing software support at the end of this month because the F330's were declared EOL a year ago. However, last I had talked to anyone about this it was believed that 5.4 would be out before they discontinued software support. What is the plans for ontap versions and F330 and support? Anyone know? ----------- Jay Orr Systems Administrator Fujitsu Nexion Inc. St. Louis, MO

1 0

IIS and netapp in existing unix world
by Mathew Anderson 14 Apr '00

14 Apr '00

Hello -- I have a 740 in production in a unix environment. I am looking forward to using this as a backend server for my NT web farm. I created a qtree and set it with mixed permissions (I also would like to set my linux web farm up and mount the same partitions). The problem that I am having is that IIS 4, (on NT4-sp6a) gives the error: Unable to enumerate files and directories because the following error occurred: Access is denied for the home directory setting, I am using a network directory and connecting to the share. I can do the a net view \\machine without a problem. An additional question is, what user should I be connecting as? and with unix permissions set for the volume (and mixed for the qtree) should there be a problem with IIS looking at the mount? I can mount the filesystem without a problem. Thanks for any help. Mathew Anderson Systems Administration NM Technet Oso Grande Technologies, Inc. mathew(a)nm.net mathew(a)osogrande.net

1 0

chown behaviour on filer NFS mounts -- help!
by Jason Ruiter 14 Apr '00

14 Apr '00

Greetings, How should chown (or fchown) behave when used on an NFS mounted filer share? I'm trying to set up an anonymous FTP server, using an NFS mounted directory as my file space. It works fine, until I try to configure anonymous uploading. Whenever I try this, I get a "fchown: not owner" error. I look at the code (wu-ftpd 2.6) and it looks like it fails on a series of fchown calls. How can I expect fchown to work? Is this OS dependent? Does the filer respond to these calls in a predictable way? I have a 720 running DataOnTap 5.3.2R1. Thank you. Jason Ruiter Systems Administrator jruiter(a)erim.org

2 1

RE: chown behaviour on filer NFS mounts -- help!
by Fox, Adam 14 Apr '00

14 Apr '00

> Greetings, > > How should chown (or fchown) behave when used on an NFS > mounted filer share? > > I'm trying to set up an anonymous FTP server, using an NFS > mounted directory > as my file space. It works fine, until I try to configure anonymous > uploading. Whenever I try this, I get a "fchown: not owner" > error. I look > at the code (wu-ftpd 2.6) and it looks like it fails on a > series of fchown > calls. > > How can I expect fchown to work? Is this OS dependent? Does the filer > respond to these calls in a predictable way? > > I have a 720 running DataOnTap 5.3.2R1. By default, only root can chown files. There is an option called wafl.root_only_chown which is set to on by default. If you turn it off and have a client that allows non-root users to chown (most modern version of UNIX do, SunOS 4.x, for example does not), then the owner of the file can chown a file as well. Now, if your ftpd is running as root, you have to be sure that the filesystem is exported with root permissions to the ftp server. Otherwise, the filer (actually any NFS server) will map root (uid 0) to nobody. Nobody will probably never own files so the process would receive a "Not Owner" message even though the process is running with an effective UID of 0. I haven't looked at the ftpd code in a while, but this is my first guess based on the info you provided. Hope this helps. -- Adam Fox NetApp Professional Services, NC adamfox(a)netapp.com

1 0

RE: Disk-to-disk backups? (was Re: Need new backup solution)
by Bond, Andrew 14 Apr '00

14 Apr '00

Hi Bruce, I'm not clear how a hacker could destroy data in snapshots - remember that they are read-only, and that's why they have proved useful in cases of virus attack. Anyone who has used tapes much will know that it's not that unusual to have problems retrieving the data, quite apart from the management of the tapes and backup software. I think the issue between use of tapes or disk for backup revolves mainly around archival. Snapshots will give you maybe a few weeks of backups, and SnapMirror can replicate this for disaster recovery, but if you need to archive for either legal reasons, or because your data is volatile (high rate of change) or needs to be kept for a long time, the amount of disk required may be too large for your computer suite, and removable media is still necessary. This issue is discussed further in a White Paper at http://www.netapp.com/tech_library/3066.html Regards, Andrew > -----Original Message----- > From: owner-dl-toasters(a)netapp.com > [mailto:owner-dl-toasters@netapp.com]On Behalf Of Bruce Sterling > Woodcock > Sent: 13 April 2000 22:58 > To: Fred Ab; toasters(a)mathworks.com; bet(a)rahul.net > Subject: Re: Disk-to-disk backups? (was Re: Need new backup solution) > > > Not only that, but if a hacker breaks in and all your backups > are online, they can just as easily destroy them too. Or they > can become corrupted due to a failure. Tapes are much more > stable (unless you pass a magnet over them) and you can > store them, as you said, offsite. > > Snapshots are great for near-term recovery, and mirrors > are even good for single-point disasters, but you always > want tapes available in the event of a catastrophe. > > Bruce > > >

1 0

How do I move /vol/vol0?
by Cliff Nadler 14 Apr '00

14 Apr '00

I'm in a situatio where we have one F760 with 4 shelves of 9Gb SCSI disk, and three shelves of 36Gb FC attached disks. The FC disks were added later, so the original install of vol0 was done to the SCSI disks. We'd like to get rid of the SCSI disks (and use one shelf of FC to hold it), but I'm concerned since we have /vol/vol0 on it. Has anyone moved /vol/vol0 off SCSI to FC before, and if so, what problems did you run into? Thanks, -- -------------------------------------------------------------------------- Cliff Nadler Collective Technologies, Inc. (a Pencom Company) cnadler(a)colltech.com (617)-443-1144 -------------------------------------------------------------------------- To win, you must treat a pressure situation as an opportunity to succeed, not an opportunity to fail Gardner Dickinson

1 0

Manual time change while using NTP...
by Eyal Traitel 14 Apr '00

14 Apr '00

Hi ! Just wanted to share with you that these steps seems to help when trying to change time manually (if you need to change time more than the timed.max_skew...) - 1. rsh toaster0 options timed.enable off 2. rsh toaster0 rdate timehost 3. rsh toaster0 options timed.enable on Pls. comment / whatever, Eyal.

2 2

Need new backup solution
by Brian Hostetter 14 Apr '00

14 Apr '00

Hi, We have been facing some explosive growth during the past 6 months and have grown past our current backup solution. We are currently backing up via a linux box that has our 740 nfs mounted. Backups are taking sometimes 24-30 hours for incrementals and all weekend for fulls. What I am looking for is info on how others are backing up 200G volumes on 740's or 760's. I just purchased a pair of 760's and am going to need a nice stable reliable backup system for them. Ideas? --Brian

8 10

Oracle Parallel Server on Filer?
by Dave Mills 14 Apr '00

14 Apr '00

Filer users, Has anyone implemented OPS on multiple NFS clients? Isn't this the product that would be required if multiple simultaneous updates to the same Oracle instance is needed? Is this a certified solution? How does file and record locking work in an OPS/Filer environment? Dave Mills Cypress Systems, LTD.

3 2

RE: Need new backup solution
by Ernst.Bokkelkamp＠mchh.siemens.de 14 Apr '00

14 Apr '00

Unfortunately this is all just theory. The reality is a bit different. You forgot that every packet on the network has to be processes and transported over multiple busses before it reaches the tape drive. These times add up to quite a bit. I have Gigabit ethernet, I have powerfull backup machines that perform at 20 Mbyte/sec on backup of ordinary Windows NT servers over LAN. These machines backup about 1.5 TB spread over 100+ servers in about 20 hours. But afte many tests the same machine with Gigabit Ethernet has only been able to backup a single filer with 4 volumes @ 139GB at a rate of about 48 GB/Hour.The Gigabit idles at that throughput. The equation that Gigabit is 10x Fast Ethernet does not work on a point-to-point situation. Bye Ernie > -----Original Message----- > From: Jay Orr [SMTP:orrjl@stl.nexen.com] > Sent: Thursday, April 13, 2000 11:22 PM > To: toasters(a)mathworks.com > Subject: RE: Need new backup solution > > On Thu, 13 Apr 2000, Schepers, Jan wrote: > > > Looking at the responses sofar, most of you agree on: > > o directly attached tape devices and > > o the use of NDMP software > > If I may put my 2 cents worth in... > > First of all as I understand NDMP (which isn't that well - feel free to > enlighten me if I'm mistaken) it's just a different standard of dump of > sorts, so other then being able to dump a netapp directly connected tape > drive there is no advantage to using it (i.e. if you're backing up over a > network, same performance of dump vs. NDMP backups). This would put the > two points above together as one. > > Second, direct connect isn't necessarly the fastest - for instance, many > tape drives have throughput of less then 10M Bytes/sec. If you have 100Mb > ethernet, thats ~11M Bytes/sec, so you're getting the same throughput and > the bottleneck is still the tape drive. If you have a seperate "backup" > network of 100Mb ethernet and backup through the network, time is saved if > you use a backup package that interleaves backups (i.e. combines backup > streams) - then the tape drive isn't waiting for one machie to feed it > data. Kick this up to Gigabit and you would outperform the tape drive. > Also, with this method you have less unused backup media and better media > inventory management. > > However, there are things that can affect throughput of networked backups > - NFS latencies, traffic on the ethernet, etc. > > SO, as I see it, there isn't really a quick-and-dirty rule of thumb on > this. The factors that would have to be considered are : > > 1) Tape throughput > 2) Network throughput > 3) Backup software methodology. > > all of these weigh close to equal in the equasion of "what's the best > way to do it", so unfortunately I don't see an easy rule of thumb.. > > Of course, I'm never one to trust an easy answer either - so maybe it's > just me ;p > > ----------- > Jay Orr > Systems Administrator > Fujitsu Nexion Inc. > St. Louis, MO

1 0

Jump to page:

2026

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