toasters October 2023

toasters@lists.teaparty.net

10 participants
5 discussions

Re: Question about flash pool maximum SSD size and local tiering
by Michael Bergman 18 Oct '23

18 Oct '23

On 2023-10-17 14:54, Florian Schmid via Toasters wrote: > Hi Johan, > thank you very much for your help. > > No, we don't have the disks yet for which the flash-pool should be used. > Not all SSDs will be used for flash-pool, only some for cache and the rest > for fast SSD storage. So you're thinking to have several different "physical tiers" with different characteristics (performance, inherent latency) for different workloads -- in the same HA-pair? Several different Aggrs with differing performance and behaviour in the same node, FAS8300? (it's a fairly powerful machine so it can do this adequately in many smaller workload cases). Or do you mean in different 8300 nodes in a X-node cluster (what's X?) This idea is much harder to make successful than you probably think. It requires you to know very much about your workloads, your applications, what they do so that you can place the correct data in the right place and you have to have the ability to do this over time as data volumes grow. Assuming they do... It's very hard indeed to automate so you need people who can baby watch this continuously and move data around. Yes that's mostly non-disruptive, but it's still quite a lot of work. It also pretty much assumes for it to be successful in the longer run that your applications do not change their workload patterns and/or pressure more than very slowly. Is this the case? All in all FabricPool is much much more automatic. It just does the job itself, pretty much w/o fuss once you've tuned it a bit w.r.t. cool down period(s) and things. It "just works". You do need an S3 target system, but as has already been pointed out it can be ONTAP with NL-SAS drives, if you already have a bunch of these lying about you can repurpose those and instead use new Cx00 (or Ax00) nodes in the "front end". The challenge with FabricPool is the network: the connection between the front end and the S3 back end needs to be very good and solid. You have to understand it fully and know every detail how it's built so you know you can trust it's capacity and latency; traffic can be quite bursty. I'm not very positive to your idea here I'm afraid: "Not all SSDs will be used for flash-pool, only some for cache and the rest for fast SSD storage." it's just my (long) experience of this that it's not very productive in reality and it costs a lot of operations (manual work, skilled personnel). It also tends to give you various problems when you need to do HW LCM (upgrade your controllers and disk back ends). It inevitably leads to stranded capacity in more than once dimension as time passes. /M -- Sr Human ;-) Alt: r.m.bergman(a)gmail-DEL_THIS-.com -- "Qui vicit non est victor nisi victus fatetur." - Ennius

3 4

Re: Re: Question about flash pool maximum SSD size and local tiering
by Gislen, Johan 17 Oct '23

17 Oct '23

Hi Florian, If you have the NL-SAS aggregate in place already I’d recommend having a look at using AWA to figure out what the optimal cache size would be: https://docs.netapp.com/us-en/ontap/disks-aggregates/determine-flash-pool-c… That would give you an idea of how much Flash Pool capacity you’d really benefit from. Regarding 3.8TB vs 7.6 TB – if the 7.6 TB drives are not listed as supported in the HWU, it’s most likely not supported. There’s also a note in the HWU for the 3.8TB drives under “System Cache Limits” – “Max number of Flash Pool Data SSDs: 20” (this is excluding RAID parity & hot spares) – if I’m intepreting that note correcly there’d be no point in adding more than one shelf of 3.8TB drives. Regards, Johan On 2023-10-17, 09:57, "Florian Schmid" <fschmid(a)ubimet.com> wrote: Hi Sebastian and Michael, thank you very much for your help. We haven't bought the SSD shelves yet. We don't know, if we should take 2 shelves with 3.8 TB or 1 shelf with 7.6 TB. I tend to one shelf with 7.6 TB... We are using flash-pool already with Raid-4, exactly for saving space. My problem is now, that I'm concerned about the HWU statement, that there is no SSD listed for using in flash-pool over 3.8 TB. This let me think, that flash-pool is not supported for SSD with 7.6 TB or greater. I haven't found anything else about it on NetApp site, only in the HWU for system cache limits for FAS 8300 Same limits are also for 8700 or 9000 series, no difference here. Really strange. Best regards, Florian ----- Ursprüngliche Mail ----- Von: "Sebastian Goetze" <spgoetze(a)gmail.com<mailto:spgoetze@gmail.com>> An: "toasters" <toasters(a)teaparty.net<mailto:toasters@teaparty.net>> Gesendet: Montag, 16. Oktober 2023 15:35:36 Betreff: Re: Question about flash pool maximum SSD size and local tiering Hi Florian, On 16.10.2023 15:01, Fenn, Michael wrote: > FlashPool SSDs (unlike FlashCache) are attached to aggregates as normal RAID groups, so you can use as many 3.8 TB drives in RAID-DP as you like to hit the maximum FlashPool capacity. Consider using RAID4 for the SSD RaidGroups... More Cache, and for the Write-cached blocks, not really less safety. (SSDs are more reliable, blocks are usually written fairly soon to HDD anyway) Regarding sizes: any size that's supported for the HW should be fine (3.8/15.3TB). Within the Cache-RG, you should use the same size disks, however... > > Note that FlashPool and FabricPool use the same underlying tiering metadata structures, so you can only have one or the other enabled on any given aggregate. > > Thanks, > Michael > > On 10/16/23, 8:03 AM, "Toasters on behalf of Florian Schmid via Toasters" <toasters-bounces(a)teaparty.net<mailto:toasters-bounces@teaparty.net> <mailto:toasters-bounces@teaparty.net<mailto:toasters-bounces@teaparty.net>> on behalf of toasters(a)teaparty.net<mailto:toasters@teaparty.net> <mailto:toasters@teaparty.net<mailto:toasters@teaparty.net>>> wrote: > > > Hi Alexander, > > this is a very good tip! Tank you very much. > I will have a look on this. > > Best regards, > Florian > > ----- Ursprüngliche Mail ----- > Von: "Alexander Griesser" <AGriesser(a)anexia.com<mailto:AGriesser@anexia.com>> > An: "Florian Schmid" <fschmid(a)ubimet.com<mailto:fschmid@ubimet.com>>, "toasters" <toasters(a)teaparty.net<mailto:toasters@teaparty.net>> > Gesendet: Montag, 16. Oktober 2023 12:07:53 > Betreff: AW: Question about flash pool maximum SSD size and local tiering > > Hi Florian, > > I cannot answer the question with the SSD sizes, I'm not sure if this is really a hard requirement or if the slices just may not be bigger than 3.8TB (in that case, you could probably manually partition the SSDs), maybe someone else has more insights into this. > > As for your second question: You can spin up OnTap's integrated S3 server on your old boxes and use them as fabric pool targets: > https://www.netapp.com/media/17219-tr4814.pdf > > Best, > > Alexander Griesser > Head of Systems Operations > > ANEXIA Internetdienstleistungs GmbH > > E-Mail: AGriesser(a)anexia.com<mailto:AGriesser@anexia.com> > Web: https://www.anexia.com > > Anschrift Hauptsitz Klagenfurt: Feldkirchnerstraße 140, 9020 Klagenfurt > Geschäftsführer: Alexander Windbichler > Firmenbuch: FN 289918a | Gerichtsstand: Klagenfurt | UID-Nummer: AT U63216601 > > -----Ursprüngliche Nachricht----- > Von: Florian Schmid <fschmid(a)ubimet.com<mailto:fschmid@ubimet.com>> > Gesendet: Montag, 16. Oktober 2023 11:53 > An: toasters(a)teaparty.net<mailto:toasters@teaparty.net> > Betreff: Question about flash pool maximum SSD size and local tiering > > Hi, > > I have checked NetApp HWU for a FAS 8300 and system cache limits. > > Ok, so far, max flash-pool is 72 TB, which is a way more than I want to use, but I haven't seen usable SSDs greater than 3.8 TB. > > Is that really true, that I can't use a 7.6 TB or 15.3 TB SSD for flash-pool? > > It would be nice, if someone has here a deeper understanding than I have about this and can give me here some clarifications. > > May I ask a second question? > Is flash-pool still the way to go for speeding up NL-SAS aggregates? > I had a look on fabric-pool tiering, but it seems like, that this only works to S3 storage, which we don't have. > We have plenty of NL-SAS storage and also of SSDs and it would be great to have a tiering between them or at least use them for caching. > > Best regards, > Florian > > > > _______________________________________________ > Toasters mailing list > Toasters(a)teaparty.net<mailto:Toasters@teaparty.net> > https://www.teaparty.net/mailman/listinfo/toasters _______________________________________________ Toasters mailing list Toasters(a)teaparty.net<mailto:Toasters@teaparty.net> https://www.teaparty.net/mailman/listinfo/toasters

2 1

Re:
by Fenn, Michael 17 Oct '23

17 Oct '23

FlashPool SSDs (unlike FlashCache) are attached to aggregates as normal RAID groups, so you can use as many 3.8 TB drives in RAID-DP as you like to hit the maximum FlashPool capacity. Note that FlashPool and FabricPool use the same underlying tiering metadata structures, so you can only have one or the other enabled on any given aggregate. Thanks, Michael On 10/16/23, 8:03 AM, "Toasters on behalf of Florian Schmid via Toasters" <toasters-bounces(a)teaparty.net <mailto:toasters-bounces@teaparty.net> on behalf of toasters(a)teaparty.net <mailto:toasters@teaparty.net>> wrote: Hi Alexander, this is a very good tip! Tank you very much. I will have a look on this. Best regards, Florian ----- Ursprüngliche Mail ----- Von: "Alexander Griesser" <AGriesser(a)anexia.com> An: "Florian Schmid" <fschmid(a)ubimet.com>, "toasters" <toasters(a)teaparty.net> Gesendet: Montag, 16. Oktober 2023 12:07:53 Betreff: AW: Question about flash pool maximum SSD size and local tiering Hi Florian, I cannot answer the question with the SSD sizes, I'm not sure if this is really a hard requirement or if the slices just may not be bigger than 3.8TB (in that case, you could probably manually partition the SSDs), maybe someone else has more insights into this. As for your second question: You can spin up OnTap's integrated S3 server on your old boxes and use them as fabric pool targets: https://www.netapp.com/media/17219-tr4814.pdf Best, Alexander Griesser Head of Systems Operations ANEXIA Internetdienstleistungs GmbH E-Mail: AGriesser(a)anexia.com Web: https://www.anexia.com Anschrift Hauptsitz Klagenfurt: Feldkirchnerstraße 140, 9020 Klagenfurt Geschäftsführer: Alexander Windbichler Firmenbuch: FN 289918a | Gerichtsstand: Klagenfurt | UID-Nummer: AT U63216601 -----Ursprüngliche Nachricht----- Von: Florian Schmid <fschmid(a)ubimet.com> Gesendet: Montag, 16. Oktober 2023 11:53 An: toasters(a)teaparty.net Betreff: Question about flash pool maximum SSD size and local tiering Hi, I have checked NetApp HWU for a FAS 8300 and system cache limits. Ok, so far, max flash-pool is 72 TB, which is a way more than I want to use, but I haven't seen usable SSDs greater than 3.8 TB. Is that really true, that I can't use a 7.6 TB or 15.3 TB SSD for flash-pool? It would be nice, if someone has here a deeper understanding than I have about this and can give me here some clarifications. May I ask a second question? Is flash-pool still the way to go for speeding up NL-SAS aggregates? I had a look on fabric-pool tiering, but it seems like, that this only works to S3 storage, which we don't have. We have plenty of NL-SAS storage and also of SSDs and it would be great to have a tiering between them or at least use them for caching. Best regards, Florian

3 2

Question about flash pool maximum SSD size and local tiering
by Florian Schmid 16 Oct '23

16 Oct '23

Hi, I have checked NetApp HWU for a FAS 8300 and system cache limits. Ok, so far, max flash-pool is 72 TB, which is a way more than I want to use, but I haven't seen usable SSDs greater than 3.8 TB. Is that really true, that I can't use a 7.6 TB or 15.3 TB SSD for flash-pool? It would be nice, if someone has here a deeper understanding than I have about this and can give me here some clarifications. May I ask a second question? Is flash-pool still the way to go for speeding up NL-SAS aggregates? I had a look on fabric-pool tiering, but it seems like, that this only works to S3 storage, which we don't have. We have plenty of NL-SAS storage and also of SSDs and it would be great to have a tiering between them or at least use them for caching. Best regards, Florian

2 2

SSH commands to ONTAP 9.11.1P11 failing from within script
by Davis, Kevin 06 Oct '23

06 Oct '23

Folks something tells me this is going to be a "D'Oh!" moment, but I'm baffled. I had a simple requirement to support an ongoing project that requires the creation of a snapshot, and a CIFS share referencing that snapshot, over a series of days. Snapshot and share name naming convention is a simple <YYYYmmdd-<string> and <YYYYmmdd-<string>$> respectively. The intention is to create both snapshot/share a little after midnight each day until the project ends and they are no longer needed. Simple enough, right? Well, when I run these commands via SSH from a prompt on a Linux box, everything works just fine: [root@foo01 ~]# /bin/ssh admin@foo'snapshot create -vserver foo3 -volume Foo04 -snapshot 20231003-foo -expiry-time 11/02/2023 00:00:00' Last login time: 10/2/2023 12:26:31 [root@foo01 ~]# /bin/ssh admin@foo 'snapshot show -vserver foo3 -volume Foo04 -snapshot 20231003-foo ' Last login time: 10/2/2023 12:46:12 Vserver: foo3 Volume: Foo04 Snapshot: 20231003-foo Creation Time: Mon Oct 02 12:46:13 2023 Snapshot Busy: false List of Owners: - Snapshot Size: 14.76MB Percentage of Total Blocks: 0% Percentage of Used Blocks: 0% Comment: - 7-Mode Snapshot: false Label for SnapMirror Operations: - Snapshot State: - Constituent Snapshot: false Expiry Time: 11/2/2023 00:00:00 SnapLock Expiry Time: - However, running a script (sh or bash, makes no difference in this case), I get this when it runs. It actually doesn't matter what ONTAP command I send, they all give the same "Vserver name: Invalid." error (the /bin/ssh lines are just echo statements in the script so I can see that it's building and sending the right command line): /bin/ssh admin@foo 'snapshot create -vserver foo3 -volume Foo04 -snapshot 20231003-foo -expiry-time 11/02/2023 00:00:00' Last login time: 10/2/2023 12:46:39 Error: Vserver name: Invalid. The Vserver name must begin with a letter or an underscore. Maximum supported length: 41 if Vserver is type "sync-source", 47 otherwise. /bin/ssh admin@foo 'cifs share create -vserver foo3 -path /Foo04/.snapshot/20231003-foo -share-name 20231003-foo$ -share-properties oplocks,browsable,changenotify,show-previous-versions -symlink-properties symlinks -offline-files manual -vscan-fileop-profile standard -max-connections-per-share 4294967295 -comment pst migrations ok to delete after 11/02/2023 00:00:00' Last login time: 10/2/2023 12:51:11 Error: Vserver name: Invalid. The Vserver name must begin with a letter or an underscore. Maximum supported length: 41 if Vserver is type "sync-source", 47 otherwise. The script is so simple it's embarrassing. Define a few variables and optargs, build 3 simple SSH commands to send to the toaster, and run them. Everything up until these commands reach the toaster is working as expected. It just breaks when it gets there. -- Kevin Davis Systems Administration | Sr. Storage Engineer | Information Services UMass Memorial Health | 100 Front St. Fl 1, Worcester, MA 01608 | mailto:kevin.davis@umassmemorial.org The information transmitted is intended only for the person or entity to which it is addressed and may contain confidential and/or privileged material. Any review, transmission, re-transmission, dissemination or other use of, or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this in error, please contact the sender and delete the material from any computer.

5 5

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 October 2023