toasters

toasters@lists.teaparty.net

13515 discussions

RE: C720 things
by niall.doherty＠netapp.com 25 Jul '00

25 Jul '00

> Thanks for that Niall, interesting.. I'm very keen to do > it, if you want to try a beta version of the software in > a production environment then let me know... We can definitely keep you in mind ! :-) Thanks ! > I know theres currently options on the netapp to allow > spoofing, Yes, the support needed on the cache server side is not too hard. Just accepting packets for us that were actually destined to a client instead ;-) (returning traffic) > but your document suggests this isnt actually workable > at the moment as the l4 switches dont redirect the return > traffic.. is that right? Woah. I said Foundry can't do it (yet), but Alteon can. We have customers doing this already. Note that you need to make sure that your traffic is "symmetrical" - i.e., the returning traffic will definitely go through the same device - otherwise you'll have problems. You have to be *very* careful with this feature... > so right now theres no way to use my cache and maintain > the client ip throughout the connection? With Alteon you can - right now. With WCCP 2 you will be able to in the (near?) future. And Foundry will hopefully add the support soon too... [ As you probably know, NetCache adds an X-Forwarded-For header (if none already exists) to all requests it sees. This contains the client IP address. So web servers can get access to it that way if they want to... ] Cheers, Niall

1 0

Monitoring other stuff with MRTG
by Henrique Pantarotto 25 Jul '00

25 Jul '00

Hello Toaster friends, I am already monitoring my filers for these infos: - CPU utilization - NFS ops/seg - /vol/vol0 volume - ethernet traffic I was wondering if its possible to also monitor: - Cache age (as reported by sysstat command) - Disk kB/s read and write (also reported by sysstat) Any know know the SNMP OID to this? I tried searching at the MIB, but I couldn't find it. Thanks, _______________________________________________ Henrique Pantarotto SysOp Site São Paulo Terra Networks Brasil S/A A Internet mais sua do que nunca Tel: (11) 5505-5728 r.316/238 ICQ: 6934285 IT: henpa henrique(a)corp.terra.com.br

1 0

RE: C720 things
by niall.doherty＠netapp.com 25 Jul '00

25 Jul '00

> On -another- thread, I'm running WCCP and of course all web > requests show the IP on the cache not the client itself.. is > there a way to spoof the client's ip? I dont have a L4 switch > btw, i run WCCP from a CAT6000 with MSFC. I have attached some info I wrote up a few days describing IP spoofing. WCCP 2 is mentioned in here. To summarise, it should be possible with WCCP 2 but we haven't tested it or added the necessary support just yet. It'll be available "soon". Coding it should be straightfoward but there are other release issues to work on first... Cheers, Niall -- Niall Doherty | Network Appliance Inc. | NetCache Deployment Architect | 495 East Java Drive | Intercontinental Operations | Sunnyvale, CA 94089, USA | mailto:ndoherty@netapp.com | http://www.netapp.com | Direct Line: +1 408 822 8554 | HQ Phone: +1 408 822 6000 | FAX(->email): +1 801 912 5212 | HQ FAX: +1 408 822 6001 |

2 1

Re: Binary numbering (off topic)
by mds＠gbnet.net 25 Jul '00

25 Jul '00

On Tue 25 Jul, 2000, Jay Orr <orrjl(a)stl.nexen.com> wrote: > > I had brought up the question a few months back of if 1GB is 1000KB or > 1024KB. Yes. Both, as the de jure and de facto standards are different. > Now as everyone knows, historically 1024 Bytes is 1KB. Seems marketing has > won the battle in terms of calling a "kilo" 1000 and not 2^10, at least > with the government. I have been involved in a thread on > comp.sys.sun.hardware about this topic and thought I'd share a posting > about this. I'm hoping I'm not the only one unaware of this "new" > standardization... Yeah, I've also heard the case presented by the faction that proposes Mibibyte, Mibibit, Kibibyte, Kibibit, etc. (sp for all those) to denote the 2^10^3n quantities. Seems quite reasonable to me.. except in the face of enormous socio-linguistic inertia I'm not sure I'd fight that hard. It's very nearly always been denary powers in raw disk sizes. I still think the binary powers are firmly ensconced in RAM, file sizes and bandwidth (MHz -> Mb and MB seems to pose less problem for people headers and preambles notwithstanding) in the mind of both the novice and the old-hand. More same-old same-old I think. I'll continue using whatever units my listeners/readers will understand most readily - anything else is poor communication. The issue that causes _me_ the most irritation is when people insist on using lowercase 'b' in Kb, Mb, etc. when they really mean 'byte' which, for me, and I suspect for many of us, would suggest 'bit'. Or when they're inconsistent. Sometimes it's not always possible to deduce from the context what they really mean, and, worse yet, they often don't see the issue as a problem! > Camiel Vanderhoeven wrote: > > > > Yes, the Ki, Mi, Gi etc. prefixes are a nice initiative to avoid > confusion. > > However, I have never seen anyone who seriously uses it. Old habits > don't > > vanish that easily. I think the simbol "KB" and the word "kilobyte" have > been > > used for 1024 bytes since the early days of computing. You can't expect > > everyone to change this into "KiB" and "kibibyte" because the SI > committe > > decides so in 1998. Not seriously. > > > > Camiel. > > > > hac <hchristeller(a)home.com> wrote: > > > > > Jay Orr wrote: > > > > > > > > I got into a discussion about a similar matter on another discussion > > > > group. Some vendors think 1 GB is 1000 MB, while others will say 1 > GB is > > > > 1024 MB. Varies by OS too. No one could is quite sure who uses > which. > > > > So, it could also be a 650MB CDROM would hold 0.634 GB.... > > > > > > > 1 GB is 1000 MB, 1 GiB is 1024 MiB. Any other use is sloppiness. > > > > > > http://physics.nist.gov/cuu/Units/binary.html > > > > > > -- > > ----------- > Jay Orr > Systems Administrator > Fujitsu Nexion Inc. > St. Louis, MO > >-- End of excerpt from Jay Orr -- -Mark ... an Englishman in London ...

1 0

Quotas for NT users
by Ferdie Mazon 25 Jul '00

25 Jul '00

Question of the day: What is the best way to use quotas in an NT environment? My user list changes frequently, and my groups are a mess. Thanks, Ferdie

1 0

Binary numbering (off topic)
by Jay Orr 25 Jul '00

25 Jul '00

I had brought up the question a few months back of if 1GB is 1000KB or 1024KB. Now as everyone knows, historically 1024 Bytes is 1KB. Seems marketing has won the battle in terms of calling a "kilo" 1000 and not 2^10, at least with the government. I have been involved in a thread on comp.sys.sun.hardware about this topic and thought I'd share a posting about this. I'm hoping I'm not the only one unaware of this "new" standardization... Camiel Vanderhoeven wrote: > > Yes, the Ki, Mi, Gi etc. prefixes are a nice initiative to avoid confusion. > However, I have never seen anyone who seriously uses it. Old habits don't > vanish that easily. I think the simbol "KB" and the word "kilobyte" have been > used for 1024 bytes since the early days of computing. You can't expect > everyone to change this into "KiB" and "kibibyte" because the SI committe > decides so in 1998. Not seriously. > > Camiel. > > hac <hchristeller(a)home.com> wrote: > > > Jay Orr wrote: > > > > > > I got into a discussion about a similar matter on another discussion > > > group. Some vendors think 1 GB is 1000 MB, while others will say 1 GB is > > > 1024 MB. Varies by OS too. No one could is quite sure who uses which. > > > So, it could also be a 650MB CDROM would hold 0.634 GB.... > > > > > 1 GB is 1000 MB, 1 GiB is 1024 MiB. Any other use is sloppiness. > > > > http://physics.nist.gov/cuu/Units/binary.html > > > > -- ----------- Jay Orr Systems Administrator Fujitsu Nexion Inc. St. Louis, MO

1 0

RE: Ndmp vs. Nfs backups.
by Kelly Wyatt 25 Jul '00

25 Jul '00

We use DLT7000, the compression devices on the netapp. Right now the master for these backups is a sun ultra 2. He really does not do much other than schedule jobs, perform database updates, etc. He has to create the list of files to backup and has to have a lot of temporary disk space and memory to handle this task. During the backups, most (all?) of the CPU cycles come from the filers -- all 760's. Here is some info on our sun: Sun Ultra 2 UPA/SBus (UltraSPARC-II 296MHz) with 384MB of memory. We have a 70GB raid array for /usr/openv to allow space for database growth. I thought he had more memory it may have been "borrowed" for an unexpected project. We do stagger our backups so that all backups jobs are not going for resources at one time. Each filer volume is a class for a lot of reasons. Originally we tried to limit our volumes to a size that can be restored within a 4 hour window... in reality, it is probably closer to 6 hours on some volumes, and we have a couple of "legacy" volumes from pre-multi volume days that are about 150GB. We have tested restores and seem to get around 15-17gb/hour restoring data, depending on the data. Kelly -- Kelly Wyatt, Kelly.Wyatt(a)SAS.com Senior Systems Programmer Strategic Recovery SAS Institute Inc. / SAS Campus Drive / Cary, NC 27513 http://www.sas.com SAS... The Power to Know -----Original Message----- From: Jay Orr [mailto:orrjl@stl.nexen.com] Sent: Tuesday, July 25, 2000 9:16 AM To: Kelly Wyatt Cc: 'Brian Hostetter'; toasters(a)mathworks.com Subject: RE: Ndmp vs. Nfs backups. Thanks for the data, it is quite useful! What kind of storage medium are you using (i.e. DLT?, AIT?) and what is your backup hardware (i.e. what are the specs of the system you use to run the backup software on)? This is the kind of info I can't even get the sales 'drioids to get for me... ----------- Jay Orr Systems Administrator Fujitsu Nexion Inc. St. Louis, MO

1 0

RE: Ndmp vs. Nfs backups.
by Kelly Wyatt 25 Jul '00

25 Jul '00

I have not tested NFS mounted backups versus NDMP direct attach (we use these in a multiprotocol manner, and with NFS mounts you lose the CIFS acls), but I have done backups to remote NDMP storage units using NetBackup. I hope this data is useful. Following is a note I sent to our management earlier this year. This is using NetBackup. The data was collected from NetBackup and thrown through SAS to perform the calculations. Kelly ----- I have been wondering what the affect of local (direct attach drive) versus remote (over the network) backups is on the NetApps. The following reports show that difference... For both reports (Full backups first report, Incr backups second report), the LOCAL is the mean GB/hour for local backups and the REMOTE is the mean GB/hour for remote backups. Where there is no REMOTE it means that there have been no remote backups for that volume. The PERDIFF is the difference between the LOCAL and the REMOTE (REMOTE/LOCAL). The anomalies (where remote is > ) probably means that the first big backup was an incremental backup, which, in reality, is equivalent to a full. Also, there are some volumes that always backup to remote filers... and this is indeed correct. This is on 100ethernet. We are looking at setting up a private GB network for remote backups. Full Backups 1 OBS CLASS LOCAL REMOTE PERDIFF 1 17.7295 8.1000 0.45687 2 b.vol0 24.9750 13.1800 0.52773 3 b.vol1 . 13.3340 . 4 h.vol0 . 12.7200 . 5 h.vol1 21.5450 12.6000 0.58482 6 h.vol2 0.5800 0.4725 0.81466 7 s.vol0 0.7200 1.8567 2.57870 8 s.vol1 15.8200 15.3133 0.96797 9 s.vol2 0.0400 0.0550 1.37500 10 s.vol3 . 0.0467 . Incremental Backups 2 OBS CLASS LOCAL REMOTE PERDIFF 1 11.6673 7.2805 0.62401 2 b.vol0 20.6550 13.6800 0.66231 3 b.vol1 11.4783 4.2700 0.37201 4 e.vol0 15.4419 12.2767 0.79502 5 e.vol1 15.6693 11.7867 0.75221 6 h.vol0 13.5387 11.0275 0.81452 7 h.vol1 19.7640 11.5750 0.58566 8 h.vol2 0.1453 0.0300 0.20642 9 k.vol0 9.2976 7.4225 0.79832 10 k.vol1 9.6065 6.9500 0.72347 11 k.vol3 10.7789 10.5400 0.97784 12 o.vol0 15.2288 9.4967 0.62360 13 o.vol1 6.7781 4.2133 0.62161 14 o.vol2 10.7219 8.6333 0.80521 15 p.vol0 1.1365 0.0700 0.06159 16 p.vol1 13.5583 9.6750 0.71358 17 p.vol2 14.8133 12.4600 0.84113 18 p.vol3 14.1555 9.8500 0.69584 19 q.vol1 0.2250 0.3300 1.46667 20 r.vol4 11.1685 11.3500 1.01625 21 s.vol1 13.4850 8.9160 0.66118 22 s.vol2 3.3287 0.0333 0.01001 23 s.vol3 0.0336 0.0200 0.59459 -- Kelly Wyatt, Kelly.Wyatt(a)SAS.com Senior Systems Programmer Strategic Recovery SAS Institute Inc. / SAS Campus Drive / Cary, NC 27513 http://www.sas.com SAS... The Power to Know -----Original Message----- From: Brian Hostetter [mailto:BrianH@dice.com] Sent: Monday, July 24, 2000 12:47 PM To: toasters(a)mathworks.com Subject: Ndmp vs. Nfs backups. Has anyone tested the speeds of ndmp backups vs nfs? I am looking at three different backup packages, and all have fuctions that will not work with ndmp (archiving, tape duplication, cataloging). I am wondering if nfs is any where near the backup speed of ndmp. If it is close then the benefits may outweigh the speed loss. Right now I am getting about ~8M/s per drive with ndmp which is really nice.

2 1

RE: Netapp examples - replication for disaster recovery?
by Anders Ljungberg S (QDT) 25 Jul '00

25 Jul '00

Hello Jim! Yes, at Ericsson in Stockholm we use this setup for disaster/recovery reasons. But we also use the replicated data to take our backups from. I.e we dont have to load our "Frontend Filers" with backup traffic. We have 10 F760's in 5 clusters replicating there volumes to 5 other F760's on another location. 5-6TB of data. We do this twice an hour and see 10-20sec bursts of data going over. Anders Ljungberg Ericsson Business Consulting -----Original Message----- From: Jim Klun 7183 [mailto:jklun@stercomm.com] Sent: Tuesday, July 25, 2000 1:37 PM To: Jim Klun 7183 Cc: Toaster Mailing List Subject: Re: Netapp examples - replication for disaster recovery? Thanks for responses - I noted that there was little mention of replication. Does anyone out there use replication filer-to-filer for disaster recovery or fail-over? I would like filers at sister sites replicating to one another to allow faster disaster recovery. Does anyone do this? * James A. Klun Sterling Commerce 4600 Lakehurst Ct. Dublin, Oh 43016 * * jklun(a)stercomm.com 614-793-7183 voice 614-793-7092 fax * * ---- "very like a whale" ---- * * ---- Hamlet, Act 2, Scene 2 ---- * On Mon, 24 Jul 2000, Jim Klun 7183 wrote: > > > I want to explain the NAS concept as implemented via Netapps in an > upcoming meeting. > > I am looking for real world examples where snapshot saved someones lunch. > Also uses of volume replication for 7X24 and DR. > > Finally - advantages - if they exist - in running Microsoft Exchange mail > off a netatpp. > > Thanks - meetings tomorrow. > > > > > * James A. Klun Sterling Commerce 4600 Lakehurst Ct. Dublin, Oh 43016 * > * jklun(a)stercomm.com 614-793-7183 voice 614-793-7092 fax * > * ---- "very like a whale" ---- * > * ---- Hamlet, Act 2, Scene 2 ---- * > >

1 0

C720 things
by Stephen J. Wilcox 25 Jul '00

25 Jul '00

Hi, it seems the C720 software has a filelimit of 2.2M objects imposed upon it, apparently netapp are due to increase this limit in a few months but in the meantime my C720 is using less than half its drive space so I want to get the max usage! At the moment I've got 51% object hit rate and 26% byte rate.. not too bad but I suspect I can increase it. As I understand it, when the file limit is hit if it retrieves a new object it replaces one of the older objects on the disk. Now, you can set caching rules up to ignore objects with certain strings in the URL eg ? or cgi-bin.. my logic is that if you get this right you can prevent the cache from storing objects which are nearly always dynamic and therefore increase the hit ratio as the objects on the drive end up being more static .. theres a similar section for non-cacheable mime-types. Theres also a section for cookie cheats to allow you to cache static objects within cookies. My question is has anyone spent time playing with these 3 fields and does anyone have some values which they have found to improve hit ratio that I can add to my (almost empty) config? On a different thread, if you add a netcache license how do you delete it, "DISABLE" doesnt work, the only way I found was to reinstall the OS on the filer. On -another- thread, I'm running WCCP and of course all web requests show the IP on the cache not the client itself.. is there a way to spoof the client's ip? I dont have a L4 switch btw, i run WCCP from a CAT6000 with MSFC. Thanks Steve -- Stephen J. Wilcox Internet Manager, Opal Telecom http://www.opaltelecom.co.uk/ Tel: 0161 222 2000 Fax: 0161 222 2008

1 0

Netapp examples needed - soon
by Jim Klun 7183 25 Jul '00

25 Jul '00

I want to explain the NAS concept as implemented via Netapps in an upcoming meeting. I am looking for real world examples where snapshot saved someones lunch. Also uses of volume replication for 7X24 and DR. Finally - advantages - if they exist - in running Microsoft Exchange mail off a netatpp. Thanks - meetings tomorrow. * James A. Klun Sterling Commerce 4600 Lakehurst Ct. Dublin, Oh 43016 * * jklun(a)stercomm.com 614-793-7183 voice 614-793-7092 fax * * ---- "very like a whale" ---- * * ---- Hamlet, Act 2, Scene 2 ---- *

2 2

Ndmp vs. Nfs backups.
by Brian Hostetter 24 Jul '00

24 Jul '00

Has anyone tested the speeds of ndmp backups vs nfs? I am looking at three different backup packages, and all have fuctions that will not work with ndmp (archiving, tape duplication, cataloging). I am wondering if nfs is any where near the backup speed of ndmp. If it is close then the benefits may outweigh the speed loss. Right now I am getting about ~8M/s per drive with ndmp which is really nice.

1 0

Cheap upgrade
by Stephen J. Wilcox 24 Jul '00

24 Jul '00

Does anyone know if there are ways to tweak extra performance out of a C720? I'm thinking about adding more RAM, can it take more and does it improve things? Any other thoughts? Thanks Steve -- Stephen J. Wilcox Internet Manager, Opal Telecommunications Plc http://www.opaltelecom.co.uk/ Tel: 0161 222 2000

3 6

Re: Linux-NFS list recommended I post here -- RE: NetApp versus Linux
by Eyal Traitel 23 Jul '00

23 Jul '00

Bennett, let's be more explicit: It takes about 2 min. to boot, probably starting to answer NFS calls a little bit earlier (it depends on your networking config - if you use trunking, it takes a few seconds to setup the trunk). The 5 minutes is if there's writing of a core file. I don't believe that generic server with big memory that crashes will take less to write its core file also (you can choose not to write a corefile, but this is probably unwise). Anyway, let's not forget that even if it is 5 minutes with core, a generic server crash will take longer to boot for replaying journaling, which the filer needs only a few seconds for. Does anyone have numbers for generic server boot times ? eT. --- Bennett Todd <bet(a)rahul.net> wrote: > 2000-07-20-16:25:44 Bruce Sterling Woodcock: > > > +++ NetApp (F720) > > > Boots in under 2 minutes due to the use of NVRAM even after an > > > improper shutdown. [...] > > > > 2 minutes is really unlikely unless you're just using vanilla NFS > > and haven't generated a core. Think more like 5 minutes (still > > impressive). > > That is such a bummer! > > I remember only a couple of years back, the proud boast was that you > could kick the plug out of a netapp, and no more than 45 seconds > after plugging it back in, it'd be live serving NFS data once again. > I even remember a salescritter saying (I think on this list) that > they'd had customers take that boast and try and feed it to 'em, > with a line like ``Ok, let's try it, if it's not back in 45 seconds > by my watch, you're outa here'', and stood by with confidence while > the box passed the test. > > When did the time increase to 2.6-6.6 times that once-proud > 45-second figure? > > At 5 minutes, that means Netapps are no longer booting significantly > faster than generic servers. > > -Bennett > > ATTACHMENT part 2 application/pgp-signature ===== Yours, Eyal Traitel eTraitel(a)yahoo.com, Home: 972-3-5290415 (Tel Aviv) *** eTraitel - it's the new eBuzzword ! *** __________________________________________________ Do You Yahoo!? Get Yahoo! Mail � Free email you can access from anywhere! http://mail.yahoo.com/

2 1

RE: very strange hang
by Francett, Robert D (SAIC) 23 Jul '00

23 Jul '00

FYI... Console activity at time of hang enclosed. Thanks, Bob Francett/BP Exploration (Alaska) <<toaster_console>> > ---------- > From: abond@netapp.com[SMTP:abond@netapp.com] > Sent: Saturday, July 22, 2000 2:38 PM > To: JenkinJp(a)bp.com; toasters(a)mathworks.com > Subject: RE: very strange hang > > Have you got a console attached? It would be interesting to see the > console > output when this occurs. > > Regards, > > Andrew Bond > Consulting Systems Engineer > Network Appliance UK > > > -----Original Message----- > > From: Jenkinson, John P (SAIC) [mailto:JenkinJp@bp.com] > > Sent: 22 July 2000 21:36 > > To: toasters(a)mathworks.com > > Subject: very strange hang > > > > > > we've got a strange one going currently > > one of 7 netapps, is a mirror mostly holding an oracle database > > several times a day, users can't authenticate to oracle or > > can't get query > > results > > nfs serving is ok > > only way outa this is to reboot the filer > > BUT > > when a reboot command is issues the files will go almost all > > the way down > > and put > > NFS Service unavailable in the filer's LCD and will need to > > be reset or > > powered off and back on to continue > > then we run for a few hours and repeat > > looked at locks and have done the sm_mon -l trick no joy > > locks may or may not be held no difference > > we do note a few errors on the quad ethernet card which is > > connected to the > > Sun Oracle server by rolled wires > > also note a few errors on the sun on the corresponding ports > > no other items look out of the norm > > > > any similar expereinces?? > > ----------------------------------------------- > > John Jenkinson > > BP Exploration (Alaska) > > >

1 0

Re: Linux-NFS list recommended I post here -- RE: NetApp versus Linux
by Eyal Traitel 22 Jul '00

22 Jul '00

I think you need to do: rc_toggle_basic panic "this is my panic message..." in order to force a core dump at the same moment. This is for strong hearts only ! eT. --- Brian Tao <taob(a)risc.org> wrote: > On Fri, 21 Jul 2000, Bennett Todd wrote: > > > > At 5 minutes, that means Netapps are no longer booting > significantly > > faster than generic servers. > > I just tried a reboot on one of our pre-production F740's, and > the > NFS client (a Solaris machine) saw an NFS outage of 1m16s. Granted, > this doesn't take into account a savecore, and I'm not sure how to > simulate that. It would be nice to have a firmware setting to > disable > the boot up memory scan and save another 10 or 15 seconds. > -- > Brian Tao (BT300, taob(a)risc.org) > "Though this be madness, yet there is method in't" > ===== Yours, Eyal Traitel eTraitel(a)yahoo.com, Home: 972-3-5290415 (Tel Aviv) *** eTraitel - it's the new eBuzzword ! *** __________________________________________________ Do You Yahoo!? Get Yahoo! Mail � Free email you can access from anywhere! http://mail.yahoo.com/

3 2

RE: very strange hang
by abond＠netapp.com 22 Jul '00

22 Jul '00

Have you got a console attached? It would be interesting to see the console output when this occurs. Regards, Andrew Bond Consulting Systems Engineer Network Appliance UK > -----Original Message----- > From: Jenkinson, John P (SAIC) [mailto:JenkinJp@bp.com] > Sent: 22 July 2000 21:36 > To: toasters(a)mathworks.com > Subject: very strange hang > > > we've got a strange one going currently > one of 7 netapps, is a mirror mostly holding an oracle database > several times a day, users can't authenticate to oracle or > can't get query > results > nfs serving is ok > only way outa this is to reboot the filer > BUT > when a reboot command is issues the files will go almost all > the way down > and put > NFS Service unavailable in the filer's LCD and will need to > be reset or > powered off and back on to continue > then we run for a few hours and repeat > looked at locks and have done the sm_mon -l trick no joy > locks may or may not be held no difference > we do note a few errors on the quad ethernet card which is > connected to the > Sun Oracle server by rolled wires > also note a few errors on the sun on the corresponding ports > no other items look out of the norm > > any similar expereinces?? > ----------------------------------------------- > John Jenkinson > BP Exploration (Alaska) >

1 0

Linux-NFS list recommended I post here -- RE: NetApp versus Linux
by Bryan -TheBS- Smith 22 Jul '00

22 Jul '00

Linux-NFS list recommended I post here -- RE: NetApp versus Linux All -- The Linux-NFS list recommended I post here. I am looking at either a $35K NetApp F720 or a $18K Linux (w/Ext3+NFS3 patches) solution. I'd greatly appreciate any input anyone has. My boss favors the NetApp solution on paper (never used one himself, nor does he do any admin-level stuff). But IMHO, unless we spend $70K to buy two (2) NetApp F720s and take advantage of the NetApp's superior failover/load-balancing capability, I'm not sure it's the right move. We still have to have a Linux server around for CVS (and NIS?) anyway, which would put a NetApp at the mercy of the Linux box anyway (any way to get CVS on NetApp?). I break down my limited knowledge below. I currently have a 100GB Linux 2.2 server (Ext2+NFS2) that I didn't setup (but have heavily reconfigured because the consulting firm didn't know jack crap, but some things like using 1KB block sizes in the /home filesystem that take forever to fsck is something I can't change right now). We're looking at going to 250GB (possibly 500-1000GB in the next 6 months). Most of our clients are Solaris 2.6 (about a dozen serious production boxes with dual to 6-way processor, 1 to 6GB of RAM, etc...). We are starting to add more and Linux workstations (both dedicated and dual-booting with NT). Stability: ---------- +++ NetApp (F720) Boots in under 2 minutes due to the use of NVRAM even after an improper shutdown. Buying two (2) units is an extremely viable failover/load-balancing solution. [ Unfortunately, my boss doesn't want to spend more than $30K). --- Linux (Ext2) Although our Linux systems have averaged >100 days uptime, a routine power outtage without proper shutdown, unable to umount NFS clients, etc... causes an fsck. We figure 30 minutes per 100GB for Ext2 fscks on 8-16KB blocks (currently takes 90 minutes of 1KB blocks, something the consultants before me did -- dumb!). Never had a problem with lost data though. --- Linux (ReiserFS) Meta-data journaling, but has NFS locking issues (except for a SuSE patch to the old knfsd?). Eliminated from consideration ??? Linux (Ext3) Full data journaling means sync I/O??? What about current stability (I'm using Trond+Higgen 0.21.3 on non-production systems with Solaris 2.6 clients and I'm happy)? Locking with Ext3 (never used Ext3, but many others are now)? I know VA Linux now has Linux 2.2 kernels with Ext3 and NFS3 that are supposively quite stable. Performance: ------------ ??? NetApp (F720 + 1000Base-SX NIC) While the NVRAM in the NetApp allows almost instant sync acknoledgement to clients, the F720 only comes with 1/4th the NVRAM of the F740/760 (8MB versus 32MB). The result is that the F720 can only handle 1/3-1/4th the NFS requests and throughput of the F740/760. I have seen good marks on the later, but cannot find much on the former which makes me wonder if we will run into a performance issue? Transmeta also mentioned that they have had performance issues with the 700-series on high-loads because the on-baord CPU, a single Alpha (usually ~600MHz) cannot even keep up with the number of XORs for just the RAID-4 volume. Of course I only have a dozen NFS clients and Transmeta has many more. Hard drives are of older, 36GB, 7200rpm, Ultra2 (aka Ultra80) SCSI type. +++ Linux (dual-cpu/memory ServerSet IIIHE chipset + dual-733MHz Pentium III + 233MHz StrongARM-powered, 4-channel, Ultra160 RAID controller on its own PCI 64-bit channel + 1000Base-SX NIC card on it's own 66MHz PCI 64-bit channel) These ServerSet III-powered mainboards seem to best both the older 4-way i450HX EDO mainboards, and the 2-way i840 in just about every memory benchmark. Going to use 2GB of RAM (which is 8x as much as in the F720), and can expand to 8GB later. The 3-independent PCI channels (of which, 2 are 66MHz PCI 64-bit) allows me to put the disk controller and network controller on their own, 64-bit channels. They Mylex 233MHz StrongARM-powered RAID controllers are the fastest you can get for just about anything (including Linux, especially when compared to vendors who still use a 100MHz i960 ASIC) and I'm going to pair it with 73GB, 10000rpm, Ultra160 disks with two disks per channel spread over all 4 channels. On the software side, I've had no issues running async with Linux. I'm more worried about Ext3's full-data journaling being the slouch here. Also wondering how long it takes, exactly, to recover after an improper shutdown with full-data journaling Ext3 (I'm assuming <5 minutes for 250GB?). [ Further insight wanted: Just wondering if the F720 will be a slouch compared to this? Or will the NVRAM really kick Ext3's butt, even if there is only 8MB of it? ] Other Considerations: --------------------- NetApp (Failover Option) Features "ready-to-go" fail-over option via proprietary gigabit links between units. Another unit can even use the disk controller and disks on a failed unit CPU unit. Very nice. Also lowers the admin requirements (like having to keep kernel/apps up-to-date, etc...). Someone on the Linux-NFS list actually mentioned administration wasn't so easy with NetApp (no web-based admin???) and would rather have Linux (I personally couldn't believe this, but she was a seasoned Linux admin like myself). Linux (Software Support) I still need a Linux box to server as the CVS server, Mail server (including various Sendmail/Mail-based programs like AV scanner, HP OpenMail, Majordomo, GNATS, etc...), NIS server (or does NetApp have NIS server capabilities?), etc... This makes me wonder if I really will "save on TOC" with the NetApp since my Linux admin outside of this has only been about 40-80 hours over the past year (compiling new kernels, installing new NFS, etc...). Cost: ----- --- NetApp F720 (7 x 36GB disk = 252GB, 216GB usable RAID-4, 180GB if using hot spare) $35K for the configuration (F720 = 600MHz? Alpha, 256MB RAM, 8MB NVRAM) with both SMB and NFS plus the 1000Base-SX upgrade. +++ Linux (5 x 73GB disk = 365GB, 292GB usable RAID-5, 219GB if using hot spare) $18K for the configuration as above (dual-733MHz, 2GB RAM, RAID controller, 1000Base-SX NIC, etc...) with SAF-TE rackmount disk chassis, ATX rackmount chassis, cabling and redundant power on both the SAF-TE and ATX case. Thanx in advance ... -- Bryan "TheBS" Smith "Lead Computer Geek" Theseus Logic, Inc. -- Bryan "TheBS" Smith CONTACT INFO *********************************************************** Chat: thebs413 @ AOL/MSN/Yahoo (see http://Everybuddy.com) Email: mailto:thebs@theseus.com,b.j.smith@ieee.org Home: http://www.SmithConcepts.com

5 8

very strange hang
by Jenkinson, John P (SAIC) 22 Jul '00

22 Jul '00

we've got a strange one going currently one of 7 netapps, is a mirror mostly holding an oracle database several times a day, users can't authenticate to oracle or can't get query results nfs serving is ok only way outa this is to reboot the filer BUT when a reboot command is issues the files will go almost all the way down and put NFS Service unavailable in the filer's LCD and will need to be reset or powered off and back on to continue then we run for a few hours and repeat looked at locks and have done the sm_mon -l trick no joy locks may or may not be held no difference we do note a few errors on the quad ethernet card which is connected to the Sun Oracle server by rolled wires also note a few errors on the sun on the corresponding ports no other items look out of the norm any similar expereinces?? ----------------------------------------------- John Jenkinson BP Exploration (Alaska)

1 0

RE: Linux-NFS list recommended I post here -- RE: NetApp versus L inux
by Graham.Knight＠netapp.com 22 Jul '00

22 Jul '00

I don't know of any way to automatically do it, but if you are at the console you can hit ESC during the memory test and it will skip it. Graham -----Original Message----- From: Brian Tao [mailto:taob@risc.org] Sent: Saturday, July 22, 2000 12:33 AM To: Bennett Todd Cc: toasters(a)mathworks.com Subject: Re: Linux-NFS list recommended I post here -- RE: NetApp versus Linux On Fri, 21 Jul 2000, Bennett Todd wrote: > > At 5 minutes, that means Netapps are no longer booting significantly > faster than generic servers. I just tried a reboot on one of our pre-production F740's, and the NFS client (a Solaris machine) saw an NFS outage of 1m16s. Granted, this doesn't take into account a savecore, and I'm not sure how to simulate that. It would be nice to have a firmware setting to disable the boot up memory scan and save another 10 or 15 seconds. -- Brian Tao (BT300, taob(a)risc.org) "Though this be madness, yet there is method in't"

1 0

Re: F760 on Ebay
by Philip Thomas 22 Jul '00

22 Jul '00

Currently $45,100.00 # of bids 36 bid history | with emails Time left Auction has ended. -- Begin original message -- > From: Fred Ab <ab(a)paloma.latticesemi.com> > Date: Fri, 21 Jul 2000 12:56:00 -0500 (CDT) > Subject: Re: F760 on Ebay > To: toasters(a)mathworks.com, busy(a)dave.web-host.net > Reply-To: Fred Ab <ab(a)paloma.latticesemi.com> > > $22,600. not $4,000. > > > ===}From: busy(a)dave.web-host.net > ===}Date: Fri, 21 Jul 2000 09:25:37 -0500 (CDT) > ===}To: toasters(a)mathworks.com > ===}Subject: F760 on Ebay > ===}Mime-Version: 1.0 > ===} > ===}All, > ===} > ===}There is an F760 filer head on ebay for about 4,000 right now. Here is > ===}the link: > ===} > ===}http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=382392343 > ===} > ===} > ===} > -- End original message -- ---philip thomas

2 1

RE: Linux-NFS list recommended I post here -- RE: NetApp versus L inux
by jeff.mohler＠netapp.com 21 Jul '00

21 Jul '00

You miss that it could take 5 minutes, in the situation of a core dump. This is a rather rare ocurrance however. -----Original Message----- From: Bennett Todd [mailto:bet@rahul.net] Sent: Friday, July 21, 2000 10:29 AM To: Bruce Sterling Woodcock Cc: toasters(a)mathworks.com Subject: Re: Linux-NFS list recommended I post here -- RE: NetApp versus Linux 2000-07-20-16:25:44 Bruce Sterling Woodcock: > > +++ NetApp (F720) > > Boots in under 2 minutes due to the use of NVRAM even after an > > improper shutdown. [...] > > 2 minutes is really unlikely unless you're just using vanilla NFS > and haven't generated a core. Think more like 5 minutes (still > impressive). That is such a bummer! I remember only a couple of years back, the proud boast was that you could kick the plug out of a netapp, and no more than 45 seconds after plugging it back in, it'd be live serving NFS data once again. I even remember a salescritter saying (I think on this list) that they'd had customers take that boast and try and feed it to 'em, with a line like ``Ok, let's try it, if it's not back in 45 seconds by my watch, you're outa here'', and stood by with confidence while the box passed the test. When did the time increase to 2.6-6.6 times that once-proud 45-second figure? At 5 minutes, that means Netapps are no longer booting significantly faster than generic servers. -Bennett

3 2

RE: Linux-NFS list recommended I post here -- RE: NetApp versus L inux
by jeff.mohler＠netapp.com 21 Jul '00

21 Jul '00

Accepted. You can probly account anywhere from 15-30 seconds for the RAM count at bootup. Why we count this..I dunno, but two minutes seems excessive with what Ive been used to in the field and as a customer. The more CIFS share you have..the longer it takes to get totally up..depends how quick your PDC responds, the more errors in your exports and hosts files (and DNS) there are will hold up the initial exportfs -a when nfs comes on, if you have an open loop FCAL adapter (one not being used) that will add a 10second delay in booting, in a failover situation the MAC address hashing thru your switched network could hold up services..etc. All things being equal, An F760 can be online serving data within a minute of a 'reboot' command WITH memory count. -----Original Message----- From: Bennett Todd [mailto:bet@rahul.net] Sent: Friday, July 21, 2000 1:09 PM To: jeff.mohler(a)netapp.com Cc: toasters(a)mathworks.com Subject: Re: Linux-NFS list recommended I post here -- RE: NetApp versus L inux 2000-07-21-13:45:38 jeff.mohler(a)netapp.com: > You miss that it could take 5 minutes, in the situation of a core dump. > > This is a rather rare ocurrance however. So 5 minutes may be rare. But Bruce also said that 2 minutes is unlikely quick. This is a radical change from the old line of 45 seconds is so reliable that account reps will happily bet a contract on it. A sad change, too. -Bennett

1 0

Re: F760 on Ebay
by Marc Nicholas 21 Jul '00

21 Jul '00

It was about $4000 when the original poster sent the email...naturally in the last few hours of the auction is rocketed. -marc On Fri, 21 Jul 2000, Fred Ab wrote: > $22,600. not $4,000. > > > ===}From: busy(a)dave.web-host.net > ===}Date: Fri, 21 Jul 2000 09:25:37 -0500 (CDT) > ===}To: toasters(a)mathworks.com > ===}Subject: F760 on Ebay > ===}Mime-Version: 1.0 > ===} > ===}All, > ===} > ===}There is an F760 filer head on ebay for about 4,000 right now. Here is > ===}the link: > ===} > ===}http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=382392343 > ===} > ===} > ===} > >

1 0

Re: RE: Re-skinning the cat, might go with the F720 wo/CIFS ... -- WA S: Linux versus NetApp
by Brian Hostetter 21 Jul '00

21 Jul '00

>>> Bryan -TheBS- Smith <thebs(a)theseus.com> 07/21 8:13 AM >>> On Thu, 20 Jul 2000, Darleen.Cormier(a)netapp.com wrote: > Bryan, You have several options, and some cost more money than others. Money is of issue. ;-> > There are several software packages that are certified with NetApps > and your library. Legato, Veritas, Workstation Solutions and > SyncSort are some of the options. Also wanted to get your opinion on Quadratec. They are a French company that has largely a European market, but Exabyte gave them a good reference for Linux, NetApp for my EZ17 + [proposed] Mammoth-2. I am testing out Quadratec's time navigator right now. So far it is pretty impressive. We had a little trouble getting it to communicate with one of our 760's. But that has been resolved and everything is happy again. Their NDMP solutions are pretty new and do not support too many media changers right now, but they tell me my qualstar 4660 will be supported soon. > These all have different configurations that they support, (you > can find the configuration info on the various web sites) > depending on where you want to put the library. All of them > support backing up the filer with the library attached locally, > but if you want to attach the library to a Solaris server, > Workstation Solutions will support that, and is also the least > expensive of the solutions listed above. Check out their website > at www.worksta.com However the cheapest solution is to attach > the library to the back of the filer, and use the dump and > restore utility. Hmmm, I'll have to check them out. Can I use Linux instead of Solaris (I'll find out, thanx for the URL ;-)? Yes, Quick Restore does work under linux, I have the cd sitting here in front of me right now. I am going to install and test it Monday. So far I have been most impressed with Backup Express from Syncsort. Their NDMP solution is great. I went from installing the CD in one of our Alpha servers to restoring files in 45 min!!! I would love to see any product that can beat that. > Several things to check out are, the library needs to be HVD > not single-ended. Hmmm, not sure I can. I guess I could get a HVD to SE converter, correct? Yes there are converters. I had to buy one from qualstar for my drives, maybe the manufacturer has one. > It needs to be set in sequential or stacker mode, That's the current mode it is in (for BRU, yuck!). The EZ17 is also set to emulate the popular EXB-210 library. > and you need to emulate an Exabyte 8900, since the filer > will not support Mammoth until the next release of software. Er, will it support it's capacity? That's crucial. Especially for Mammoth-2. Thanx for all your help and insight. -- TheBS -- Bryan "TheBS" Smith CONTACT INFO *********************************************************** Chat: thebs413 @ AOL/MSN/Yahoo (see http://Everybuddy.com) Email: mailto:thebs@theseus.com,b.j.smith@ieee.org Legal: http://www.SmithConcepts.com/legal.html

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