toasters

toasters@lists.teaparty.net

13531 discussions

NetApp File Server (FS230) mounted directly on an NT server
by Weeks, Thomas 18 May '98

18 May '98

I have a fairly good grasp on NT and CIFS and the like... but I am from the UNIX worls and just want someone to answer this for me once and all... Is there any way to MOUNT a NetApp file server directly to an NT file system and access it like unix does with a /mnt directory? I know about the menthod of mounting to NT via a drive letter... but you can't re-share that as a resource under that the NT file system CAN you? You know.. like this: given that NetApp box="\\barium\C$" and NT Server="\\ntbox" net use \\ntbox\shares\share \\barium\share net use \\ntbox\shares\winapps \\barium\winapps I know this is probably not possible.. but ya never know if you dont' ask.. it just seems that if the netapp box could be treated as a raw storage unit like this, that it would greatly simplify matters (regarding all the ACL problems). Anyone wanna tell me exactly why things can't be done like this? Thomas Weeks GTE SYSTEMS ENGINEER Testing-Integration Center "The Pit" Brooks AFB, DSN 240-5444

1 0

Exports file vs exportfs
by Alexei Rodriguez 18 May '98

18 May '98

In 4.x, NetApp added a feature such that one could do an export of a file system to a network (net/netmask) such that: exportfs -root=192.168.250.0/24 /export/test would allow anybody on the 192.168.250 network to get to /export/test on the filer. Cool. This works without a problem. What I am running into a hitch with is making this work out of the /etc/exports file. It just does not seem to want to work. Has anybody else worked with this notation out of the exports file? Did it work? Is the order of the hosts important (subnets before hosts)? Or am I simply missing something obvious? Thanks. Alex

1 1

Re: Support for Network Time Protocol?
by Jason D. Kelleher 18 May '98

18 May '98

In message <199805162033.NAA20194(a)paradiso.cnet.com>, John Caruso writes: >> Nothing religious about it, actually. We have a well established NTP >> infrastructure, we've used NTP for many years and are very comfortable >> with it. Thus we'd like to use it for NetApps too. > >Ditto. This seems to be on the "sometime in the future..." list for >Netapp, but we haven't heard of any definite plans. I recall asking about NTP over two years ago. Very disappointing. jason --- Jason D. Kelleher kelleher(a)susq.com Susquehanna Investment Group 610.617.2721 (voice) 401 City Line Ave, Suite 220 610.617.2916 (fax) Bala Cynwyd, PA 19004-1122

1 0

NetApp as database storage - interface speeds?
by Thomson, Michael 18 May '98

18 May '98

Hi there. I've been asked to look at whether NetApp is a potential candidate for hosting some Oracle databases, but one thing I don't see is how to get sufficient sustained data rates between the Oracle server and the filer. We use 10 Base-T Ethernet and FDDI right now, 100 Base-TX Ethernet is coming onstream at present. Even connecting a filer direct to a server on private 100 Base-TX only gives 10 MBytes/sec, lots slower than Ultra-SCSI. If I'm going to use NetApp to host Oracle Databases I probably want more than one database server connected to a filer. I can only have one Gigabit Ethernet card per F630. So right now, I can't see a way to support more than one Oracle server from a single NetApp. Yours comments much appreciated as always. Cheers, Mike Thomson

2 1

Re: Support for Network Time Protocol?
by sirbruce＠ix.netcom.com 18 May '98

18 May '98

On 05/16/98 08:52:28 you wrote: > >> It is a little bit of a hassle, but rsh rdate works pretty well. Any reason >> why you don't want to use it? Do you use NTP for the rest of your environment? > >Yes. NTP for everything, including our Cisco routers. Keeps the time nicely, >and doesn't step it. It would be *much* nicer to have NetApp support use >NTP also. Okay, I was just curious. Most places I've seen aren't as "religious" about the use of NTP. Not that there's anything wrong with that. :) Bruce

2 1

Re: Support for Network Time Protocol?
by Alan Judge 18 May '98

18 May '98

Ditto here. We use NTP for everything else, machines, routers, even some Windoze PCs, ... It gives us accurate multi-machine logs that we can stand over (important for an ISP when people are accused of various sorts of misbehaviour). We also asked Netapp about NTP, but haven't had any definite statement yet. Even SNTP would do. -- Alan

1 0

HELP! CIFS and 4.3.4?
by Weeks, Thomas 18 May '98

18 May '98

Hey all... I was told, by just about everyone, that if I wanted full CIFS without the "backout problems" possible with 5.x, then I should just go to 4.3.4... Well.. I have.. and the major porblem I HAD with the NetApp implimentation of CIFS seems to still be there: It looks like while yes there is indeed NT server authentication happening now, no matter WHO I want to access the CIFS shares, they STILL have to MANUALLY PLACED into the /etc/passwd file!!!! What is the POINT of full blown NT server (domain) authentication if I STILL have to input ALL of the users into the /etc/passwd file! That's what my main problem is! We have TONS of UNIX users across campus who are on various NIS+ domains and as a result many have overlapping UIDs! i.e. we can NOT rely on single NIS+ uid mapping.. and would have the SAME problem if we tried to input all our users directly into a single /etc/passwd file on the server. Is the ONLY way to get around this is by going to 5.x? Does 5.x even FIX this serious problem? What about the differences in the 5.x file system? I don't HAVE a test file server to test this on.. 100% production... 8v/ Anyone here have this kind of expereince? Thomas Weeks GTE SYSTEMS ENGINEER Testing-Integration Center "The Pit" Brooks AFB, DSN 240-5444

2 1

Re: Performance problems
by sirbruce＠ix.netcom.com 18 May '98

18 May '98

On 05/14/98 15:19:12 you wrote: > > >How are the NICs on the F520 configured? How are the >switch ports configured? > >Make sure that the duplexes match on both ends. And, after checking that, I would double-check your setup in terms of routing. Perhaps configurations and netstat -r for each BSD system and filer would be helpful. Bruce

3 2

Re: Support for Network Time Protocol?
by sirbruce＠ix.netcom.com 17 May '98

17 May '98

On 05/15/98 14:39:03 you wrote: > >We recently observed some strange behavior due to the filer's clock being >about 1 minute and 30 seconds faster than a client's. The dates displayed >by 'ls -l' of files that had been recently created were listed as "May 14 >1998" instead of "May 14 18:22" like they should be. That is odd... I though ls did this only if the file was older than six months. I presume after 90 seconds it displays normally again, so I guess the ls in question isn't handling "in the future" times very well (it thinks they are in the past). >It is important that NFS clients and servers have the same time. We would >prefer to use NTP for time synchronization on the NetApp instead of running >a cron job on a UNIX server that will fire off a 'rsh rdate' command on the >filer. It is a little bit of a hassle, but rsh rdate works pretty well. Any reason why you don't want to use it? Do you use NTP for the rest of your environment? >Does anyone know about any NetApp plans to support NTP? I've heard a lot of people ask for it over the years but usually they move on to a different solution since HTP doesn't seem immediately forthcoming. Bruce

2 1

Re: Botched filer upgrade
by sirbruce＠ix.netcom.com 17 May '98

17 May '98

On 05/16/98 15:05:47 you wrote: > > >Hey folks. > >I just got some gigabit ethernet cards to replace my fddi >back-end network. The install went well (trivial) but all >release, non-5.x, DataONTAP versions have some bugs with >the gigabit ethernet. > >So, I upgraded to 4.3.4D2 which supposedly has the fix. >It works well on my f540. > >My f630 pukes horribly: > >Probing devices >Finding image... >Loading 14 f 5 disk@0 >Complete >Starting NetApp Release 4.3.4D2: Wed May 13 07:55:56 PDT 1998 >Copyright (c) 1992-1998 Network Appliance, Inc. > >PANIC: MM fault MMCSR=1, type=1, VA=0x79dba0, PC=0xfffffc00002a2d44 > > >version: NetApp Release 4.3.4D2: Wed May 13 07:55:56 PDT 1998 >cc flags: 1 >dumping core: Error dumping core --- not dumped. > >ok c > > >I quickly put the 4.3.3 boot floppy I had and we are back in service >and I am calming down.. :) > >So, anybody able to decode this error? It's a pretty brutal error that you probably can't do anything to fix. Since 4.3.4D2 is unreleased and a patch you got from support (I assume) particularly to address a Gigabit problem, then my guess it was never tested on a 600 series. So you really need to talk to tech support and let them know of the error and ask when a version of that will be available for the 600 series. :) Bruce

1 0

Re: Support for Network Time Protocol?
by sthaug＠nethelp.no 16 May '98

16 May '98

> >Yes. NTP for everything, including our Cisco routers. Keeps the time nicely, > >and doesn't step it. It would be *much* nicer to have NetApp support use > >NTP also. > > Okay, I was just curious. Most places I've seen aren't as "religious" about > the use of NTP. Not that there's anything wrong with that. :) Nothing religious about it, actually. We have a well established NTP infrastructure, we've used NTP for many years and are very comfortable with it. Thus we'd like to use it for NetApps too. Steinar Haug, Nethelp consulting, sthaug(a)nethelp.no

2 1

Botched filer upgrade
by Alexei Rodriguez 16 May '98

16 May '98

Hey folks. I just got some gigabit ethernet cards to replace my fddi back-end network. The install went well (trivial) but all release, non-5.x, DataONTAP versions have some bugs with the gigabit ethernet. So, I upgraded to 4.3.4D2 which supposedly has the fix. It works well on my f540. My f630 pukes horribly: Probing devices Finding image... Loading 14 f 5 disk@0 Complete Starting NetApp Release 4.3.4D2: Wed May 13 07:55:56 PDT 1998 Copyright (c) 1992-1998 Network Appliance, Inc. PANIC: MM fault MMCSR=1, type=1, VA=0x79dba0, PC=0xfffffc00002a2d44 version: NetApp Release 4.3.4D2: Wed May 13 07:55:56 PDT 1998 cc flags: 1 dumping core: Error dumping core --- not dumped. ok c I quickly put the 4.3.3 boot floppy I had and we are back in service and I am calming down.. :) So, anybody able to decode this error? My filer is: whirlpool> sysconfig -v NetApp Release 4.3.3: Mon Mar 16 12:58:27 PST 1998 System ID: 0016786960 slot 0: System Board (NetApp System Board IV rev-b) Firmware release: 1.8_a1 Memory Size: 512 MB slot 2: Ethernet Controller e2 (Znyx) MAC Address: 00:c0:95:f8:2f:6d (Twisted pair) slot 3: FDDI Network Adapter f3 (DEC 3002) MAC Address: 00:00:f8:bd:c6:04 (Dual Attach Fiber) slot 8: Dual SCSI Host Adapter (NetApp SCSI Adapter I) SCSI Host Adapter 8a (Qlogic ISP 1040B) Firmware Version 2.26 Clock Rate 40MHz. 0: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 1: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 2: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 3: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 8: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 9: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 10: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 11: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 12: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) EMU 0: P/N 500408 S/N 2550270236 Firmware Rev 1.19 0 Swap Events, 0 Enclosure Events EMU 1: P/N 500408 S/N 2550272530 Firmware Rev 1.19 0 Swap Events, 0 Enclosure Events SCSI Host Adapter 8b (Qlogic ISP 1040B) Firmware Version 2.26 Clock Rate 40MHz. 0: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 1: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 2: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 3: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 9: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 10: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 11: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) 12: SEAGATE ST19171W 0023 Size=8.3GB (17783112 blocks) EMU 2: P/N 500408 S/N 2550272276 Firmware Rev 1.19 0 Swap Events, 0 Enclosure Events EMU 3: P/N 500408 S/N 2550272259 Firmware Rev 1.19 0 Swap Events, 0 Enclosure Events slot 9: NVRAM (NetApp NVRAM II) Revision: A1 Serial Number: 9744 Memory Size: 16 MB Battery 1 Status: 100% (3.30v) Battery 2 Status: 100% (3.29v) slot 10: Gigabit Ethernet Controller e10 MAC Address: 00:60:cf:20:48:61 (1000fx) whirlpool> I need gigabit pretty badly but don't want to go to 5.x yet. Anybody? Thanks. Alex

1 0

Re: Support for Network Time Protocol?
by sthaug＠nethelp.no 16 May '98

16 May '98

> It is a little bit of a hassle, but rsh rdate works pretty well. Any reason > why you don't want to use it? Do you use NTP for the rest of your environment? Yes. NTP for everything, including our Cisco routers. Keeps the time nicely, and doesn't step it. It would be *much* nicer to have NetApp support use NTP also. Steinar Haug, Nethelp consulting, sthaug(a)nethelp.no

1 0

Re: Calculating increase in used space when moving to a filer (long)
by Keith Brown 15 May '98

15 May '98

>I'm about to move data from an old&creaking NT server to a F230. >Iwas aware that moving to the filer would lead to a space usage >increase due to the larger block size (4K as against 512bytes for >standard NTFS) This is not necessarily the case, and almost certainly wouldn't be the case for most installations/migrations. "Standard NTFS" would very *rarely* be using an allocation unit size of 512 bytes on most modern systems. The way NTFS figures out the default allocation unit size to use on a particular disk or volume-set is quite simple, if you can remember the "kilobyte to gigabyte" correlation that exists. Specifically, NTFS uses an allocation unit size of half a kilobyte on disks of half a gigabyte or less (rare these days, right? :-)), one kilobyte on disks of one gigabyte or less, two kilobytes on disks of 2 gigabytes or less, and 4 kilobytes thereafter. Unless you manually created your NTFS file systems using something along the lines of a: C:\> FORMAT d: /fs:ntfs /a:512 to override the default and force an allocation unit size of 512 bytes, then your NTFS file system will likely be using the same default allocation unit size as WAFL uses (4Kb). Also, don't forget the swings-and-roundabouts issues that are normally associated with small AU sizes. There's a reason that AU sizes generally get bigger as file system sizes increase. Smallers AU's cause most file systems to burn more AU's to keep track of the AU's, if you see what I'm saying? :-) It is often hard at best, and impossible at worst to accurately account for such matters in application level scripts that just read "visible stuff", but I don't think that you'll find that WAFL and NTFS are fantastically different space burners for the vast majority of situations.* If you do, well.... let me know, *especially* if WAFL proves far superior.... :-) Keith

1 0

Support for Network Time Protocol?
by Mike Mueller 15 May '98

15 May '98

We recently observed some strange behavior due to the filer's clock being about 1 minute and 30 seconds faster than a client's. The dates displayed by 'ls -l' of files that had been recently created were listed as "May 14 1998" instead of "May 14 18:22" like they should be. It is important that NFS clients and servers have the same time. We would prefer to use NTP for time synchronization on the NetApp instead of running a cron job on a UNIX server that will fire off a 'rsh rdate' command on the filer. Does anyone know about any NetApp plans to support NTP? -- Mike

1 0

Re: syslog and /etc/messages format
by tjt＠netapp.com 15 May '98

15 May '98

> for one reason. There isn't an appendix which contains a _complete_ > list of all error messages and sample _syntax_ in the filer > documentation. ... > So, I'm all for more specific messages -- as long as they are > documented. Please. Thanks for the input. Yes, it's our intention that unique message ids be usable for looking up information about the messages, implying that there is a database of all messages that can used in a variety of ways, including documentation. ...Tim Thompson...Network Appliance, Inc...tjt(a)netapp.com...

1 0

Re: Performance problems
by Devi Nagaraj 15 May '98

15 May '98

How are the NICs on the F520 configured? How are the switch ports configured? Make sure that the duplexes match on both ends. -Devi At 04:09 PM 5/14/98 -0500, Brett Rabe wrote: > >Most of ya'll probably understand what it feels >like to be frustratingly baffled, so pull up >a chair, lend me an ear, and I'll buy ya a beer. > >I've got two NetApps (and two more on order). > >1) F230 > o 5.0R2 OnTap > o 128MB RAM > o 4MB NVRAM > o 5 9-gig drives > o 2 100bT NIC > >2) F520 > o 5.0R2 OnTap > o 192MB RAM > o 4 MB NVRAM > o 9 9-gig drives > o 2 100bT NIC > >The installation is as follows. Both NetApps >are fronted by three FreeBSD servers. One NIC >on the NetApps routes to its own private subnet. >Each of the three servers have a second NIC card >also routing to that subnet. > >I can't get squat for performance from the F520. >I'm not trying to pretend that using 'yes' is a >true benchmark, but it does illustrate the point >sufficiently: > >F520 front-end# yes > /some_nfs_mounted_filesystem/yes >F520> sysstat 1 > CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache > in out read write read write age > 2% 35 0 0 445 8 0 0 0 0 >60 > 4% 36 0 0 468 9 0 0 0 0 >60 > 18% 37 0 0 469 9 1684 816 0 0 >60 > 10% 30 0 0 377 7 96 1632 0 0 >60 > 15% 157 0 0 1457 33 0 0 0 0 >60 > 29% 84 0 0 870 18 1336 2348 0 0 >60 > > >F230 front-end# yes > /some_nfs_mounted_filesystem/yes >F230> sysstat 1 > CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache > in out read write read write age > 68% 287 0 0 1510 823 1052 2300 0 0 >60 > 63% 212 0 0 1683 751 24 1628 0 0 >60 > 67% 248 0 0 1678 811 944 1960 0 0 >60 > 59% 248 0 0 1680 780 1068 1632 0 0 >60 > 59% 214 0 0 1727 804 1048 1760 0 0 >60 > 60% 215 0 0 1684 805 43 2285 0 0 >60 > 53% 202 0 0 1741 783 1052 2328 0 0 >60 > 57% 179 0 0 1416 758 1064 2320 0 0 >60 > >(The F230 also was being backed up and is in >production hosting web pages, in case anyone >notes about all the outbound traffic that wouldn't >be accounted for by 'yes'....) > >The NetApps are configured as close to identical >as reasonable. The version of FreeBSD appears to >be not a factor -- I've got both 2.2.5-RELEASE and >2.2.6-RELEASE that produce the second results on >the F230, and 2.2.5-RELEASE and 2.2.6-RELEASE that >produce the first results on the F520. > >All the interfaces are configured correctly. There >are no significant errors reported by the NetApps, >the FreeBSD boxes, or the network device to which >they all terminate. I can't discern any significant >difference in the FreeBSD box configurations -- if >anything, the fact that I get the same performance >with different FreeBSD OS revs on the F230 and >the same poor performance with different FreeBSD OS >revs on the F520 points to a problem with the F520. > >I had the F230 at 4.3.1 and got similar performance >as I do with 5.0R2. I've gone so far as to swap >ethernet cables. :p > >I'm not sure where to look next; suggestions would >be highly welcomed. I have (of course) hit now.netapp.com >and hunted through their troubleshooting suggestions >to no avail. > >... just another day. :) > >Brett > >--- >Brett Rabe Email : brett(a)uswest.net >Systems Administrator - U S West Phone : 612.664.3078 >Interact - 3S Pager : 612.613.2549 >600 Stinson Blvd. Pager : page-brett(a)uswest.net >Minneapolis, MN USA Fax : 612.664.4770 > > All those who believe in psychokinesis, raise my hand! > > >

4 5

Calculating increase in used space when moving to a filer (long)
by Garrett Burke 15 May '98

15 May '98

Hi All, I'm about to move data from an old&creaking NT server to a F230. I was aware that moving to the filer would lead to a space usage increase due to the larger block size (4K as against 512bytes for standard NTFS) Having searched on NOW, ToolTips and on this list, I bit the bullet and wrote my own script to calculate the space increase. Problem is that a 'du' on a FreeBSD box shows more blocks being used on the filer than I calculated. It's close, but I'd be interested in finding out what the difference is. Someone else must have wrote a script like mine to calculate the change. Is the script missing something (an earlier version missed the space allocated to the directories themselves) or is the filer allocating space for something else. How many directory entries can a 4K directory hold? (There's a total of 337 directories in what I'm running the script on) I have a very deep hierarchy to move. Here's what the script produces when run on a NT 4 SP 3 with a mapped share from a NT 4 SP 3 server. H:\ta\1.60\ NT FS size 216115.695313 KB (3720 files) Netapp FS size 223892.000000 (56311 blocks) KB increase 7776.304688 KB (3%) FreeBSD opc007.msc.ie 2.2-BETA_A FreeBSD (I know, I know, but it's a production box that was up for 80 days before the power went off and on and off and on again) NFS mounted F230 running 4.3.3 opc007:1011#pwd /toaster/releases/ta/1.60 opc007:1012#set | grep BLOCK BLOCKSIZE=4096 opc007:1013#du -s 56505 . opc007:1014# There's a difference of 194, which isn't fixed. It changes for a different hierarchy. The script is below. I know it's not very efficient, but it does run on both NT and UNIX. Use it (Correct it!) if you wish, but at your own risk. (As an aside I'd be interested in hearing from other filer users that are running CIFS, especially in a SW dev environment.) Regards, GB. #!/usr/bin/perl # # Calculate the space used on a filer with 4K block sizes for # the files in the current directory & it`s sub directories # # Based on the example on pages 56/57 of Programming Perl 1st Ed # by Larry Wall and Randal L. Schwartz pub O'Reilly # # Use at your own risk # # Garrett.Burke(a)msc.ie 15/5/98 $blk=4096; # no. of bytes per netapp block &dodir('.'); sub dodir{ local($dir,$nlink)=@_; local($dev,$ino,$mode,$subcount); local(@allfiles); print "Working in $dir\n"; opendir(DIR,'.') || die "Can't open $dir"; @allfiles=readdir(DIR); closedir(DIR); for (@allfiles) { if ($_ eq '.') { # Don`t forget the directories themselves $numblks++; next; } next if $_ eq '..'; $name="$dir/$_"; $fnums++; ($dev,$ino,$mode,$nlink,$uid,$gif,$rdev,$size, $atime,$mtime,$ctime,$blksize,$blocks)=stat($_); if (-d $_) { # print "$name is a directory\n"; chdir $_ || die "Can't cd to $name"; &dodir($name); chdir ('..'); } else { $numfiles++; $nt+=$size/1024; # NT size in KBs (fp) $netappblks=int($size / $blk); # no. of (int)netapp blocks if ($size % $blk) { $netappblks++; # anything over then another block } $numblks+=$netappblks; $netappfs+=$netappblks*4; # FS size (fp) in KB on netapp } } # for @ } if ($nt) { $increase=$netappfs-$nt; # fp printf("NT FS size %f KB (%d files) Netapp FS size %f (%d blocks) KB increase %f KB (%d\%) \n", $nt,$numfiles,$netappfs,$numblks,$increase,($increase/$nt)*100); } -- Garrett Burke, Network Admin,Managed Solutions Corp. 32 Upper Mount St., Dublin, Ireland. burkeg(a)msc.ie +353-1-661-4840 A jug of wine, a loaf of bread and thou.

1 0

Re: Performance problems
by sirbruce＠ix.netcom.com 15 May '98

15 May '98

On 05/14/98 18:56:07 you wrote: > >> What, by the way, is "yes"? > > >YES(1) FreeBSD General Commands Manual YES(1) [etc.] An old UNIX hand once told me yes was originally "designed" for use with the fsck -y option. Whether or not this was true, I don't know. Other things to note - on some systems it's still a C program, on others it's just a shell script. Also, I think some shells have "yes" built-in now. So this can also account for performance differences between OSes. Bruce

2 1

syslog and /etc/messages format
by tjt＠netapp.com 15 May '98

15 May '98

Here at NetApp, we're planning some changes in the way messages are implemented, to provide more specific information (e.g. unique message id's, to identify exactly where messages are coming from), handle things like internationalization, and make it easier to post-process messages. We'd like to know how many people currently have scripts to process the /etc/messages file (and/or any other destinations of syslog or console output). If you would be affected by any changes to the format of /etc/messages or console output, please send me email, so we can take your input into account. Thanks much. ...Tim Thompson...Network Appliance, Inc...tjt(a)netapp.com...

2 1

Performance problems
by Brett Rabe 14 May '98

14 May '98

Most of ya'll probably understand what it feels like to be frustratingly baffled, so pull up a chair, lend me an ear, and I'll buy ya a beer. I've got two NetApps (and two more on order). 1) F230 o 5.0R2 OnTap o 128MB RAM o 4MB NVRAM o 5 9-gig drives o 2 100bT NIC 2) F520 o 5.0R2 OnTap o 192MB RAM o 4 MB NVRAM o 9 9-gig drives o 2 100bT NIC The installation is as follows. Both NetApps are fronted by three FreeBSD servers. One NIC on the NetApps routes to its own private subnet. Each of the three servers have a second NIC card also routing to that subnet. I can't get squat for performance from the F520. I'm not trying to pretend that using 'yes' is a true benchmark, but it does illustrate the point sufficiently: F520 front-end# yes > /some_nfs_mounted_filesystem/yes F520> sysstat 1 CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache in out read write read write age 2% 35 0 0 445 8 0 0 0 0 >60 4% 36 0 0 468 9 0 0 0 0 >60 18% 37 0 0 469 9 1684 816 0 0 >60 10% 30 0 0 377 7 96 1632 0 0 >60 15% 157 0 0 1457 33 0 0 0 0 >60 29% 84 0 0 870 18 1336 2348 0 0 >60 F230 front-end# yes > /some_nfs_mounted_filesystem/yes F230> sysstat 1 CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache in out read write read write age 68% 287 0 0 1510 823 1052 2300 0 0 >60 63% 212 0 0 1683 751 24 1628 0 0 >60 67% 248 0 0 1678 811 944 1960 0 0 >60 59% 248 0 0 1680 780 1068 1632 0 0 >60 59% 214 0 0 1727 804 1048 1760 0 0 >60 60% 215 0 0 1684 805 43 2285 0 0 >60 53% 202 0 0 1741 783 1052 2328 0 0 >60 57% 179 0 0 1416 758 1064 2320 0 0 >60 (The F230 also was being backed up and is in production hosting web pages, in case anyone notes about all the outbound traffic that wouldn't be accounted for by 'yes'....) The NetApps are configured as close to identical as reasonable. The version of FreeBSD appears to be not a factor -- I've got both 2.2.5-RELEASE and 2.2.6-RELEASE that produce the second results on the F230, and 2.2.5-RELEASE and 2.2.6-RELEASE that produce the first results on the F520. All the interfaces are configured correctly. There are no significant errors reported by the NetApps, the FreeBSD boxes, or the network device to which they all terminate. I can't discern any significant difference in the FreeBSD box configurations -- if anything, the fact that I get the same performance with different FreeBSD OS revs on the F230 and the same poor performance with different FreeBSD OS revs on the F520 points to a problem with the F520. I had the F230 at 4.3.1 and got similar performance as I do with 5.0R2. I've gone so far as to swap ethernet cables. :p I'm not sure where to look next; suggestions would be highly welcomed. I have (of course) hit now.netapp.com and hunted through their troubleshooting suggestions to no avail. ... just another day. :) Brett --- Brett Rabe Email : brett(a)uswest.net Systems Administrator - U S West Phone : 612.664.3078 Interact - 3S Pager : 612.613.2549 600 Stinson Blvd. Pager : page-brett(a)uswest.net Minneapolis, MN USA Fax : 612.664.4770 All those who believe in psychokinesis, raise my hand!

1 0

Earthlink's NetApp-based email system?
by mer＠world.evansville.net 13 May '98

13 May '98

A few months ago I read an article on NetApp's webserver which was written by an Earthlink engineer. The article described the work that Earthlink had done to build a scalable, reliable email system using NetApp filers. It was interesting to me because it covered in detail the work needed to ensure proper locking across NFS for this application. The article seems to have been removed from www.netapp.com, and none of the search engines I have tried know of it anymore either. Does anyone have an URL (or a copy) they could send me? Thanks, -- Marc Rouleau VP and Chief Technology Officer Voice: (812) 479-1700 Fax: (812) 479-3439 Network WCS http://www.networkwcs.com

2 1

new filer and netcache monitoring tools on NOW site
by tjt＠netapp.com 13 May '98

13 May '98

The Tools&Utilities section of the NetApp NOW site (http://now.netapp.com) has a couple of new tools for monitoring: filer-mrtg A package that uses MRTG for monitoring filers. It provides a web-based interface for current and historical views of: network traffic in/out, NFS ops, disk usage (%), CPU usage, CIFS read/write ops, and CIFS other ops. netcache-mrtg A package that uses MRTG for monitoring NetCache Appliances. It provides a web-based interface for current and historical views of: URLs served, Cache I/O and hit rates, network bandwidth, URL fetch time, cpu and disk space usage, and DNS calls. If you haven't checked out the Tools&Utilities section of NOW lately, it has been enhanced with a number of new tools, and a "Tool Time" effort in which customers are encourage to submit tools (and win a fabulous polo shirt!). ...Tim Thompson...Network Appliance, Inc...tjt(a)netapp.com...

1 0

subscribe
by jimy＠hwacomsys.com.tw 12 May '98

12 May '98

1 0

Re: Toasters list - any FAQs?
by sirbruce＠ix.netcom.com 08 May '98

08 May '98

On 05/08/98 21:19:41 you wrote: > >> Also, here's a real item: I'm seeing fairly high >> error and collision rates on my F330's quad ethernet >> card, (10BaseT Quad-Ethernet Controller (Znyx)) and >> was wondering if there are any known issues with this >> card, since the network traffic on these four nets >> does not warrent this high error rate (I'm on OnTap 4.3.1). > >Note that collision is not the same as error! On standard half-duplex >Ethernet, a collision is a normal and healthy phenomenon - simply means >that the Ethernet flow control meachnism is working. > >It's okay to be worried about high error rates. High *collision* rates >(as long as we're not talking about "late collision") aren't really all >that meaningfull by themselves - what you should probably look at is the >*utilization* of the Ethernet segment. I myself have observed high error rates on the Quad 10BaseT card on the filer. It was clearly the largest source of errors on the network, and much higher than I'd ever seen on a normal network. Moreover, it seemed that the error rate increased which each port (I don't quite remember if it went from A to D or D to A) without correlation to the actual amount of traffic on those ports. A bug was filed on this. However, these errors never seemed to actually be a major problem for the network and while I suppose they were responsible for the occasional performance anomaly overall performance of the filer was never a problem. Bruce

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