toasters November 2016

toasters@lists.teaparty.net

19 participants
8 discussions

Flash Cache vs Flash Pool
by Ehrenwald, Ian 21 Dec '16

21 Dec '16

Hello Toasters I'm thinking about if/how to implement Flash Pool in our new cDOT clusters, and was wondering if anyone could provide a bit of real world guidance for me. We have two SAS aggregates, aggr_sas_600g_c1n1 with 8x24 600g and aggr_sas_1200g_c1n1 with 4x24 1.2t. I've been doing a bunch of reading about Flash Pool vs Flash Cache and am trying to better understand their strengths and weaknesses. Flash Pool accelerates writes as well as reads (Flash Cache is reads only), however with Flash Pool there seems to be the potential for slower cache access/throughput vs Flash Cache since the data needs to travel the SAS path vs Flash Cache which is probably DMA through PCIe. Maybe that's not a concern at all, I don't know. Additionally, it appears that using Flash Pool disables the Flash Cache functionality for the aggregates which are in hybrid mode (makes sense), but then we have expensive add-in cards doing nothing. Our theoretical Flash Pool would be 2x24 200g, giving us about 5.5t of usable caching space to sprinkle into these aggregates. I've been running AWA on the cluster against those two SAS aggregates for ~24 hours and have come up with these stats: ### FP AWA Stats ### Host mrk_c1n1 Memory 61054 MB ONTAP Version NetApp Release 8.3.2P5: Tue Aug 23 01:27:00 PDT 2016 Basic Information Aggregate aggr_sas_600g_c1n1 Current-time Tue Nov 22 17:06:53 EST 2016 Start-time Mon Nov 21 12:48:17 EST 2016 Total runtime (sec) 101918 Interval length (sec) 600 Total intervals 157 In-core Intervals 1024 Summary of the past 157 intervals max ------------ Read Throughput (MB/s): 339.039 Write Throughput (MB/s): 123.536 Cacheable Read (%): 56 Cacheable Write (%): 66 Max Projected Cache Size (GiB): 787.463 Summary Cache Hit Rate vs. Cache Size Referenced Cache Size (GiB): 714.650 Referenced Interval: ID 132 starting at Tue Nov 22 12:33:05 EST 2016 Size 20% 40% 60% 80% 100% Read Hit (%) 25 30 30 30 33 Write Hit (%) 1 2 2 2 2 The entire results and output of Automated Workload Analyzer (AWA) are estimates. The format, syntax, CLI, results and output of AWA may change in future Data ONTAP releases. AWA reports the projected cache size in capacity. It does not make recommendations regarding the number of data SSDs required. Please follow the guidelines for configuring and deploying Flash Pool; that are provided in tools and collateral documents. These include verifying the platform cache size maximums and minimum number and maximum number of data SSDs. Basic Information Aggregate aggr_sas_1200g_c1n1 Current-time Tue Nov 22 17:06:53 EST 2016 Start-time Mon Nov 21 12:40:57 EST 2016 Total runtime (sec) 102357 Interval length (sec) 600 Total intervals 158 In-core Intervals 1024 Summary of the past 158 intervals max ------------ Read Throughput (MB/s): 914.247 Write Throughput (MB/s): 257.318 Cacheable Read (%): 41 Cacheable Write (%): 26 Max Projected Cache Size (GiB): 2412.178 Summary Cache Hit Rate vs. Cache Size Referenced Cache Size (GiB): 2142.380 Referenced Interval: ID 113 starting at Tue Nov 22 09:04:41 EST 2016 Size 20% 40% 60% 80% 100% Read Hit (%) 34 38 38 38 41 Write Hit (%) 7 7 7 7 9 The entire results and output of Automated Workload Analyzer (AWA) are estimates. The format, syntax, CLI, results and output of AWA may change in future Data ONTAP releases. AWA reports the projected cache size in capacity. It does not make recommendations regarding the number of data SSDs required. Please follow the guidelines for configuring and deploying Flash Pool; that are provided in tools and collateral documents. These include verifying the platform cache size maximums and minimum number and maximum number of data SSDs. ### FP AWA Stats End ### Aggregate aggr_sas_600g_c1n1 has a lot of random overwrites (66%) that could have been cached. The volumes in that aggregate are pretty much exclusively Oracle databases. The other aggregate, aggr_sas_1200g_c1n1, doesn't seem hit as hard. Given those statistics, what would you do if your options were ~5.5t of Flash Pool vs buying another 2t Flash Cache card per node in this HA pair? I seem to be missing the 'Projected Read Offload' and 'Projected Write Offload' statistics which would have been very useful, mentioned at the Flash Pool documentation in https://library.netapp.com/ecmdocs/ECMP1368404/html/GUID-2C3EC0DF-FEFE-4871 -A161-4A3BAC87DB69.html Thanks for any insight you all can provide. -- Ian Ehrenwald Senior Infrastructure Engineer Hachette Book Group, Inc. 1.617.263.1948 / ian.ehrenwald(a)hbgusa.com This may contain confidential material. If you are not an intended recipient, please notify the sender, delete immediately, and understand that no disclosure or reliance on the information herein is permitted. Hachette Book Group may monitor email to and from our network.

4 8

FDMI reporting with qla2xxx driver on Linux
by Momonth 09 Dec '16

09 Dec '16

Hi, I recently noticed that qla2xxx (Qlogic HBAs) driver on Linux is capable of reporting FDMI. By default it sends "Linux-default", and I'm failing to see how to change it to actual host name as it becomes way more handy then =) Any ideas? Cheers, Vladimir

2 3

ONTAP9 NFSv4 + CentOS 6.8 does not work
by Momonth 29 Nov '16

29 Nov '16

Hi All, We upgraded our 6 nodes cluster to ONTAP9 yesterday. Today I discovered some freshly installed CentOS 6.8 hosts not being able to mount file shares via NFS v4 protocol. The mount command just kind of hangs. I verified it's not a connectivity issue, ie connections to the filer's port 2049 are possible. Also NFS v3 works just fine. Did anyone experience anything like that? Cheers, Vladimir

7 11

CDOT snapmirror -forground
by Rhodes, Richard L. 28 Nov '16

28 Nov '16

Hello I am starting to look into a possible rollover of our 7-mode IBM N-Series systems to new NetApp CDOT systems. I run backups with snapmirror and snapvault that are controlled by a AIX ksh script. The ksh script runs the following cmds to perform the data transfers: snapmirror update -S ${filerpri}:${vol} -w ${filerdr}:${vol} snapvault snap create -w $vol $vol Note the '-w' parm. I run these cmds in forground, so all backups are single threaded and return a response back to the script. I'm looking at CDOT cmds manual for an equivalent '-w' parm. I find the following for the "snapmirror update" cmd: [-foreground | -w [true]] - Foreground Process This specifies whether the operation runs as a foreground process. If this parameter is specified, the default setting is true (the operation runs in the foreground). When set to true, the command will not return until the process completes. This parameter is only applicable to relationships with "Relationship Capability" of "Pre 8.2". Q) What does "applicable to relationships with "Relationship Capability" of "Pre 8.2" mean? Does this mean that a CDOT>8.2 system doesn't support this parm? Thanks! Rick ----------------------------------------- The information contained in this message is intended only for the personal and confidential use of the recipient(s) named above. If the reader of this message is not the intended recipient or an agent responsible for delivering it to the intended recipient, you are hereby notified that you have received this document in error and that any review, dissemination, distribution, or copying of this message is strictly prohibited. If you have received this communication in error, please notify us immediately, and delete the original message.

2 1

Flash Cache older 3020c??
by Chris hansen 26 Nov '16

26 Nov '16

I saw a few threads talking about flash cache and it made me think of my old 3020c that continues to chug along in my lab running 7.3.7. (568 days) Do any of these PAM II flash modules (111-00525+D0 for example) work in this FAS model or was there no flash option? I did some googling but there wasn’t much data on some of these older modules. -Chris

2 1

NFS V4 and nobody
by John Adams 16 Nov '16

16 Nov '16

I am mounting an export via NFS V4. At first, I saw that mounted clients showed user/group as nobody:nobody. But then I configured the nfs.v4.id.domain option on the filer to be the client machines DNS domain name. Then the username:groups were correct.. mostly... We use NIS on the client machines and the filer. For files that are owned by users or groups that are not in our NIS, the clients now shows nobody. Under NFS V3... this shows the UID/GID numbers (if they are not in NIS, and have no mappings). Anybody know how to make NFS V4 just show the UID/GID of entries that have no mapping instead of converting them to nobody? (UID's are no longer in our NIS after the account is deleted...) RHEL6/7. OnTap 8.2.4-7 Mode. Thanks.

3 3

Problems with SMB and macOS Sierra?
by James Andrewartha 09 Nov '16

09 Nov '16

Hi, We're testing macOS Sierra at the moment, and are seeing some weird problems connecting to SMB shares on our 8.2.4 7-mode filer. Listing directories is super-slow, and then after a while all the files get the no permissions icon and you can't open anything. Has anyone seen anything similar? Thanks, -- James Andrewartha Network & Projects Engineer Christ Church Grammar School Claremont, Western Australia Ph. (08) 9442 1757 Mob. 0424 160 877

4 7

Troubleshooting 'unreachable' on a CDOT port
by Edward Rolison 08 Nov '16

08 Nov '16

Hello fellow NetApp admins. Over the weekend, I hit a really quite "interesting" sort of a problem. One of those weekends that ... no one really wants to have. It involved a firmware upgrade on a switch going catastrophically wrong. causing chaos for several hours. And off the back of that - one of our two CDOT nodes, it's primary 'data' interface LACP group has .. seemingly died. I say "seemingly" because: - Snooping the interfaces sees packets going in and out. (Mostly "arp"). - But the switch side "snoop" doesn't see the ARP replies. And thus never 'learns' the mac, and doesn't route traffic to it. This happens on both ports of a LACP group, and even moving it to another switch entirely hasn't helped. Manually offlining the ports doesn't help (Except if I do both, it migrates the lif automatically). But switching the lifs over onto the other head - has fixed it, for now. (although obviously, we're failed over, and have reduced resilience). Has anyone run into anything similar? Or can give me some insight as to what could explain this perplexing behaviour? The 'source' of the problem was - probably - some serious network strangeness. Loops, vlans going up and down, all sorts of chaos. I haven't (yet) rebooted the failed node, as the vservers are running quite merrily. I'm wondering if there's some sort of DOS/arp flood protection that might be tripping us up. Thanks, Ed.

4 4

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 November 2016