Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
It may be a tiny bit more complicated, however, after you administer it for a short while you will find it is very close to 7-mode.
I now prefer it to 7-mode due to many enhancements (too many to list mind you).
7-mode will be around, just do not expect any new features in 7-mode. The effort is going into Clustered Data Ontap ( aka cDOT)
With 8.2 and the release of a single-node "cluster", for new installations, there is really no reason to use 7-mode.
Existing installations should eventually migrate, but the effort will be worth the while.
--tmac
*Tim McCarthy* *Principal Consultant*
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141https://www.redhat.com/wapps/training/certification/verify.html?certNumber=110-107-141&isSearch=False&verify=Verify NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 3:25 PM, Iluhes iluhes@yahoo.com wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle?
That must have been a while ago since SAN has been supported in cDOT starting with 8.1.
Regarding migration from 7-mode to cDOT, NetApp has a tool named 7MTT that leverages SnapMirror to migrate data from 7-mode to cDOT filers.
-k
On Nov 6, 2013, at 12:50 PM, "John Stoffel" john@stoffel.org wrote:
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle? _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
The tool appears to be directed to small installations. If you have large/complex systems, it would be best to enlist the NetApp PS crew.
I used the the tool that preceeded the 7MTT...the Volume Transition Wizard...which by the way may panic a destination running 8.2...found out the hard way! The VTW worked OK prior to 8.2.
It was easy to setup per volume. I have not played much at all with the new tool as now that I am migrated, I really have no use for it ;)
--tmac
*Tim McCarthy* *Principal Consultant*
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141https://www.redhat.com/wapps/training/certification/verify.html?certNumber=110-107-141&isSearch=False&verify=Verify NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 4:02 PM, Koji Kabuto kabutovr6@yahoo.com wrote:
That must have been a while ago since SAN has been supported in cDOT starting with 8.1.
Regarding migration from 7-mode to cDOT, NetApp has a tool named 7MTT that leverages SnapMirror to migrate data from 7-mode to cDOT filers.
-k
On Nov 6, 2013, at 12:50 PM, "John Stoffel" john@stoffel.org wrote:
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle? _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Hello
With classic or 7G Data ONTAP, all of the administration is stored in flat files on each storage system.
Using Clustered Data ONTAP, all of the administrative information is stored in relational databases that are shared by all systems in the cluster.
This feels more elegant to me because it means that if I want to create a user to perform a function such as backup operator, in 7G, I had to do so on every storage system, in cDOT, I create the user once for all the systems.
The same holds true for what we used to call exports and now call export policies. You create them once per cluster even if you have 20 storage systems in a cluster.
And what is really handy and nice is the tab for completion. Once you get used to tabbing through a command line, you want to use that on your 7-Mode systems.
But you might feel like it is necessary when simple command like: ping nodename becomes network ping -node local -destination node name
Ramp up takes some time.
In my opinion, Clustered Data ONTAP is a paradigm shift. You can simple map the commands one to one between 7G and cDOT because the implications are different but once you get used to it, it is very hard to go back. You feel like you are traveling back to the early 90s when you use the 7G systems.
That is my take on it.
--April
On Wednesday, November 6, 2013 1:15 PM, tmac tmacmd@gmail.com wrote:
The tool appears to be directed to small installations. If you have large/complex systems, it would be best to enlist the NetApp PS crew.
I used the the tool that preceeded the 7MTT...the Volume Transition Wizard...which by the way may panic a destination running 8.2...found out the hard way! The VTW worked OK prior to 8.2.
It was easy to setup per volume. I have not played much at all with the new tool as now that I am migrated, I really have no use for it ;)
--tmac
Tim McCarthy Principal Consultant
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141 NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 4:02 PM, Koji Kabuto kabutovr6@yahoo.com wrote:
That must have been a while ago since SAN has been supported in cDOT starting with 8.1.
Regarding migration from 7-mode to cDOT, NetApp has a tool named 7MTT that leverages SnapMirror to migrate data from 7-mode to cDOT filers.
-k
On Nov 6, 2013, at 12:50 PM, "John Stoffel" john@stoffel.org wrote:
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle?
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
_______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Now that you can snapmirror from 7-mode to cDOT, see vid below demonstrating it, it's one less hurdle:
http://netapp.tynsoe.org/wordpress/7-mode-cdot-screencat/
I have read that 8.2 7-mode is the last development release, all subsequent releases of 7-mode will be maintenance releases. This is further confirmed when you compare the 7-mode and cDOT release notes and see how many features are in cDOT and not 7-mode e.g NFS4, SMB3, QOS.
I also noticed that the next version (v6) of Oncommand Core (aka DFM) is cDOT only. A big issue for us is that cDOT doesn't have the concept of vFiler DR but I think there are more efficient ways of doing DR at the application level.
I don't think Netapp will need to discontinue 7-mode, as times goes on it will lack an increasing number of features and Customers will migrate to cDOT or another vendor.
-- View this message in context: http://network-appliance-toasters.10978.n7.nabble.com/Cluster-DATA-vs-7-mode... Sent from the Network Appliance - Toasters mailing list archive at Nabble.com.
From what I see, both the 3220/4350 support FC targets.
SO, I am not sure about your comment. I use SAN on my nodes. You can set up one or more vServers (very similar, but slightly different than a vFiler) to serve FC/iSCSI LUNS and/or NFS/CIFS traffic
In my case (I am a partner and did my own Professional Services work), I was able to procure a new FAS3170 pair with a TON of disks (enough to split betwixt 4 other heads!)
I staged my data one or a few volume(s) at a time. WIth tools, I was able to use the equivalent of SnapMirror to send traffic to the new Clustered ONTAP. After the data was migrated we updated maps and pointed to the new locations.
It took about three months, but I had something like 400 volumes to move. Got it done.
Your milage will vary. Bottom line, it would likely be best to secure the use of NetApp Professional Services to aid in the migration effort if you have a large/complicated installation.
--tmac
*Tim McCarthy* *Principal Consultant*
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141https://www.redhat.com/wapps/training/certification/verify.html?certNumber=110-107-141&isSearch=False&verify=Verify NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 3:50 PM, John Stoffel john@stoffel.org wrote:
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle?
Hi John,
just for the record, you're mixing up vFilers (7-Mode / MultiStore) and vServers (clustered ONTAP)...
cDOT vServer support FC SAN since 8.1, as others have already mentioned.
One thing holding back people here in Europe is lack of MetroCluster support. Once that is fixed I expect Europe to catch up with the cDOT adoption rate.
Sebastian
On 11/6/2013 9:50 PM, John Stoffel wrote:
tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle? _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
From what I have heard we should expect to see MetroCluster in 8.3.
-k
On Nov 6, 2013, at 6:11 PM, Sebastian Goetze spgoetze@gmail.com wrote:
Hi John,
just for the record, you're mixing up vFilers (7-Mode / MultiStore) and vServers (clustered ONTAP)...
cDOT vServer support FC SAN since 8.1, as others have already mentioned.
One thing holding back people here in Europe is lack of MetroCluster support. Once that is fixed I expect Europe to catch up with the cDOT adoption rate.
Sebastian
On 11/6/2013 9:50 PM, John Stoffel wrote: tmac> It may be a tiny bit more complicated, however, after you tmac> administer it for a short while you will find it is very close tmac> to 7-mode.
tmac> I now prefer it to 7-mode due to many enhancements (too many to tmac> list mind you).
But what about the limitations? We looked at it when we moved up to 3220s and 3250 pairs, but in particular the vFilers not being able to do SAN FC LUNs was a big turn off for us. I wanted to setup a certain group in a vFiler, but since they are FC only mostly, it was a no-go.
We're currently running 8.1.2 in 7-mode and fairly happy. Thinking up upgrades, but if it ain't seriously broke, why bother?
Also, how would we migrate from 7-mode to cluster mode with a minimum of hassle? _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
On 07/11/13 08:11, Sebastian Goetze wrote:
One thing holding back people here in Europe is lack of MetroCluster support. Once that is fixed I expect Europe to catch up with the cDOT adoption rate.
Speaking of MetroCluster, did anyone else notice the launch of Optical SAS? It allows the use of single-mode fibre for disks in stretch metros, although FCVI still needs to be over multimode. You also save from not having to use the 6500N FC-SAS bridges.
Other limitations on Optical SAS:
- If optical SAS cables are used for inter-shelf cables, optical SAS cables must also be used to connect to storage controller - Only one pair of patch panels can be used between connections - Only 6Gb/sec configurations are currently supported, no IOM3 support - No support for DS4486 in MetroCluster configuration - Limitation of 96 SSD spindles per stack - FAS22xx NOT supported
https://communities.netapp.com/community/netapp-blogs/netapp-360/blog/2013/1...
http://www.netapp.com/us/media/sas-storage-cabling-and-infrastructure.pdf
I would say you have no choice.. DFM 5.x for 7 mode is only going to be bug fixes.. I would say there is a slight learning curve, but if you download the map between 7 and CM, it helps.. Being able to make wildcard changes and auto tab is really cool.. They really cleaned up the commands, and now you have a global name space.. NAS is really the ticket, and SAN is kinda limited to 4 x4 nodes.. Scaling to more then 2 nodes is what other storage vendors will aspire to as they reach memory and CPU constraints (Isolon has some similar scale features/issues).. More can chime in on the block side, but you have lots of cool options, especially on the network side of the house.. Only new features such has CIFS 3.0 is only supported in CM..
Just my opinion..
Date: Wed, 6 Nov 2013 12:25:17 -0800 From: iluhes@yahoo.com Subject: Cluster DATA vs 7-mode: admin prospective To: toasters@teaparty.net
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS? _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John
What is a site like ours to do?
We have a pair of primary filers (FAS3220) and a single backup filer (FAS2240). We don't have any spare kit lying around which would run 8.2 (our older off-support filers are all too old for that).
The filers are <1 year old, we went with 7 mode because of the lack of support for Qtrees and SnapVault in 8.1.2 c-mode which was the newest available at installation.
Am I stuck on 7 mode until we do our next technology refresh in ~ 5 years?
Jeremy
On 07/11/2013, at 9:01 AM, John Clear wrote:
On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John _______________________________________________ Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
On 07/11/13 06:17, Jeremy Webber wrote:
We have a pair of primary filers (FAS3220) and a single backup filer (FAS2240). We don't have any spare kit lying around which would run 8.2 (our older off-support filers are all too old for that).
The filers are <1 year old, we went with 7 mode because of the lack of support for Qtrees and SnapVault in 8.1.2 c-mode which was the newest available at installation.
Am I stuck on 7 mode until we do our next technology refresh in ~ 5 years?
My guess is yes. We have a nearly identical setup to yours (installed 2011 with 5 years support), and I was recently recommended to move VMware backups from OnCommand 5 to VSC because OC 5 Host Package won't see any bug fixes now OC6 is cDOT only.
On 20/11/2013, at 3:05 PM, James Andrewartha wrote:
On 07/11/13 06:17, Jeremy Webber wrote:
Am I stuck on 7 mode until we do our next technology refresh in ~ 5 years?
My guess is yes. We have a nearly identical setup to yours (installed 2011 with 5 years support), and I was recently recommended to move VMware backups from OnCommand 5 to VSC because OC 5 Host Package won't see any bug fixes now OC6 is cDOT only.
NetApp I am very very unhappy about this. We bought expensive, enterprise class kit and only 1 year after purchase it is on care-and-maintenance by the vendor? (Yes I'll be taking it up with our local sales reps next time I see them)
I had been researching using VSC vs OnCommand to back up vSphere previously - I found the restore facilities in OnCommand to be primitive beyond belief so I would have gone the VSC route (which at least can sometimes do file level restores) in any case.
Jeremy
On 21/11/13 10:27, Jeremy Webber wrote:
NetApp I am very very unhappy about this. We bought expensive, enterprise class kit and only 1 year after purchase it is on care-and-maintenance by the vendor? (Yes I'll be taking it up with our local sales reps next time I see them)
Yeah, I think they have been too quick to drop 7-mode development as well, particularly given how much one pays in support.
I had been researching using VSC vs OnCommand to back up vSphere previously - I found the restore facilities in OnCommand to be primitive beyond belief so I would have gone the VSC route (which at least can sometimes do file level restores) in any case.
Don't worry, Single File Restore is being removed from the next version of VSC as well: https://communities.netapp.com/community/products_and_solutions/virtualizati...
I can think of many things that would clearly annoy me:
1. Not having flat files to manage and quickly modify, quotas: exports, rc, Running config commands potentially much more dangerous then 2. RSH? Can I run RSH commands? 3. No separate exports for qtrees 4. Not clear how to access log file of the cluster....mounting root share not possible, right? only for vServer.. 5. No local user authentication mode for cifs vServer 6. Complicated model of Logical interfaces spreading around away from vServer 7. ".admin" in NFS mount path ....Common... 8. Anti-virus?
On Wednesday, November 6, 2013 4:01 PM, John Clear jac@panix.com wrote:
On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John
If you are so worried...download the Ontap Simulator and try it out. ALl those items you listed are not really any big deal.
--tmac
*Tim McCarthy* *Principal Consultant*
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141https://www.redhat.com/wapps/training/certification/verify.html?certNumber=110-107-141&isSearch=False&verify=Verify NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 6:37 PM, Iluhes iluhes@yahoo.com wrote:
I can think of many things that would clearly annoy me:
- Not having flat files to manage and quickly modify, quotas: exports,
rc, Running config commands potentially much more dangerous then 2. RSH? Can I run RSH commands? 3. No separate exports for qtrees 4. Not clear how to access log file of the cluster....mounting root share not possible, right? only for vServer.. 5. No local user authentication mode for cifs vServer 6. Complicated model of Logical interfaces spreading around away from vServer 7. ".admin" in NFS mount path ....Common... 8. Anti-virus?
On Wednesday, November 6, 2013 4:01 PM, John Clear jac@panix.com wrote: On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
On 11/06/2013 03:37 PM, Iluhes wrote:
I can think of many things that would clearly annoy me:
- Not having flat files to manage and quickly modify, quotas: exports,
rc, Running config commands potentially much more dangerous then
bothered me at first, too, but the CLI is rich, and the API is getting there. Also take a look at Work Flow Automator as a way to do repetitive tasks.
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
- No local user authentication mode for cifs vServer
We use AD. I suspect its on the roadmap, though, as I see a cifs option called "Is Local Auth Enabled" that sounds intriguing.
- Complicated model of Logical interfaces spreading around away from
vServer
not been an issue for us.
- ".admin" in NFS mount path ....Common...
how is this an issue?
- Anti-virus?
soon.
FWIW, we've been a cDOT (formerly known as "GX") user since 2006. Initially GX was put in-place for scaleout HPC file service, and its stayed around in that role ever since. Now that the data protection features (snap mirror, snap vault) are available, we're migrating to as much cDOT as possible.
Why?
Vol move, aggregate relocate, LIF migrate and the fundamental ability to do all system currency tasks, resource balancing, add and remove hardware and OnTap upgrades without service disruption.
With our 7-Mode fleet, when a controller is saturated, a downtime is needed to move a volume to another controller or aggregate. Its a short-downtime usually, but still requires stopping file service access to that data set.
With cDOT, I can do every one of the common operational and systems currency tasks we do, without disrupting clients.
Totally worth it.
As for migration, our model is to migrate as Systems ReCap and projects start, over the next year or two. There will be service interruptions as data sets move from 7-mode to cDOT, but once in, there should be no disruptions for access to that data set for the rest of its lifetime.
-skottie
On Wednesday, November 6, 2013 4:01 PM, John Clear jac@panix.com wrote: On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
- Complicated model of Logical interfaces spreading around away from
vServer
Right. really a non-issue. In fact, you can impliment on-board load-balancing between lifs. Really handy. Turn on NFSv4 and PNFS and now you even get direct data path. (all nodes handle meta-data, redirects client to most efficeint data path) In 7-mode, you still have a interface for data per node. Same in cDOT. You just have a LOT more flexibility on how to use them. Need to do maintenance? Migrate the lif to another node. I just did this to turn on jumbo frames and test. Works great.
--tmac
*Tim McCarthy* *Principal Consultant*
Clustered ONTAP Clustered ONTAP NCDA ID: XK7R3GEKC1QQ2LVD RHCE6 110-107-141https://www.redhat.com/wapps/training/certification/verify.html?certNumber=110-107-141&isSearch=False&verify=Verify NCSIE ID: C14QPHE21FR4YWD4 Expires: 08 November 2014 Current until Aug 02, 2016 Expires: 08 November 2014
On Wed, Nov 6, 2013 at 7:54 PM, Scott Miller Scott.Miller@dreamworks.comwrote:
On 11/06/2013 03:37 PM, Iluhes wrote:
I can think of many things that would clearly annoy me:
- Not having flat files to manage and quickly modify, quotas: exports,
rc, Running config commands potentially much more dangerous then
bothered me at first, too, but the CLI is rich, and the API is getting there. Also take a look at Work Flow Automator as a way to do repetitive tasks.
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
- No local user authentication mode for cifs vServer
We use AD. I suspect its on the roadmap, though, as I see a cifs option called "Is Local Auth Enabled" that sounds intriguing.
- Complicated model of Logical interfaces spreading around away from
vServer
not been an issue for us.
- ".admin" in NFS mount path ....Common...
how is this an issue?
- Anti-virus?
soon.
FWIW, we've been a cDOT (formerly known as "GX") user since 2006. Initially GX was put in-place for scaleout HPC file service, and its stayed around in that role ever since. Now that the data protection features (snap mirror, snap vault) are available, we're migrating to as much cDOT as possible.
Why?
Vol move, aggregate relocate, LIF migrate and the fundamental ability to do all system currency tasks, resource balancing, add and remove hardware and OnTap upgrades without service disruption.
With our 7-Mode fleet, when a controller is saturated, a downtime is needed to move a volume to another controller or aggregate. Its a short-downtime usually, but still requires stopping file service access to that data set.
With cDOT, I can do every one of the common operational and systems currency tasks we do, without disrupting clients.
Totally worth it.
As for migration, our model is to migrate as Systems ReCap and projects start, over the next year or two. There will be service interruptions as data sets move from 7-mode to cDOT, but once in, there should be no disruptions for access to that data set for the rest of its lifetime.
-skottie
On Wednesday, November 6, 2013 4:01 PM, John Clear jac@panix.com wrote: On Wed, Nov 06, 2013 at 12:25:17PM -0800, Iluhes wrote:
Netapp is pushing hard to start using Cluster Data ONTAP. From what I see, from admin/operator level is not as elegant and friendly as "Classic". I just cannot see going to it, it goes against almost everything I like about administrating netapp. Do people feel the same way out there is the field? Netapp is telling us that 7-mode will go away. Is that for real? Do you "like" Cluster OS?
Prior to 8.2, we were only implementing CDOT where we needed the extra performance of all the nodes appearing as a single system, or the ability to migrate volumes between nodes.
With the release of 8.2, all the features of 7-mode that we use are supported in CDOT. The main one for us was Snapvault, but QOS is also a plus. We are not installing any more 7-mode filers.
CDOT is still a bit rough around the edges, but administering it in some ways is easier than 7-mode. When I have to work on our 7-mode systems now, I find myself missing tab completion, wildcards and global commands.
YMMV,
John
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Just adding a few things:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys) See KB: 1012524 https://kb.netapp.com/support/index?page=content&id=1012542&actp=sea...
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ...” commands
You can also enable remost access so that you can view the logs from http. This is very useful.
See: https://kb.netapp.com/support/index?page=content&id=1013814
Cheers,
—April
To summarize: To enable the web services for a cluster, use these steps from the KB article mentioned above: 1. Enable the required Web services for the cluster: cluster-1::> vserver services web modify -name spi|ontapi|compat -vserver * -enabled true 2. Create the Web service access service spi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name spi -role admin -vserver cluster1 3. Create the Web service access service ontapi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name ontapi -role admin -vserver cluster1 Note: This error will be returned for the admin role, as it is pre-created on cluster create. This error will not be returned if a different role name that is not pre-created is provided. Error: command failed: duplicate entry 4. Create the Web service access service compat for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name compat -role admin -vserver cluster1 5. Identify the cluster-mgmt LIF for the cluster. This will be used for HTTPS access. A node-mgmt LIF can also be used instead of the cluster-mgmt LIF: cluster-1::*> net int show -role cluster-mgmt -fields address (network interface show) vserver lif address -------- ----------- ----------- cluster-1 clus_mgmt 10.98.243.120 6. Identify the names of the nodes in the cluster: cluster-1:> node show -fields node node ---------- cluster1-01 cluster1-02 Cluster1-03 Cluster1-04 4 entries were displayed.
6. Add the login to allow http access. cluster1::> login create -username admin -application http -authmethod password
cluster1::> login show -username admin -application http (security login show) Authentication Acct Vserver UserName Application Method Role Name Locked ----------- ---------------- ----------- -------------- ---------------- ------ cluster1 admin http password admin no
The following URLs will provide access to the log directory on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/log/ https://10.98.243.120/spi/node-a/etc/log/ https://10.98.243.120/spi/node-b/etc/log/ https://10.98.243.120/spi/node-c/etc/log/ https://10.98.243.120/spi/node-d/etc/log/
The following URLs will provide access to the directory where the core files are saved on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/crash/ https://10.98.243.120/spi/node-a/etc/crash/ https://10.98.243.120/spi/node-b/etc/crash/ https://10.98.243.120/spi/node-c/etc/crash/ https://10.98.243.120/spi/node-d/etc/crash/
On Nov 6, 2013, at 4:54 PM, Scott Miller Scott.Miller@dreamworks.com wrote:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
I guess coming from unix background, I always found very useful having ability to have an admin box with all root volumes mounted and just search/grep messages file, copy and edit RC files, rsh scripts via scripts for quota reports, backups, etc. I am sure there might be a way to get some/most/all of that functionality with cDOT, but just the elegance of administering everything from single linux admin hosts just does not seem to be present or fully capable anymore. Am I wrong? I also never worked on anything that need to grow beyond 2-nodes, single cluster. Why would netapp force cluster on such setups, I would have think that majority of install base is small, medium, 2 node clusters
On Wednesday, November 6, 2013 8:08 PM, aprilogi aprilogi@yahoo.com wrote:
Just adding a few things:
2. RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys) See KB: 1012524 https://kb.netapp.com/support/index?page=content&id=1012542&actp=sea...
4. Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ...” commands
You can also enable remost access so that you can view the logs from http. This is very useful.
See: https://kb.netapp.com/support/index?page=content&id=1013814
Cheers,
—April
To summarize: To enable the webservices for a cluster, use these steps from the KB article mentioned above: 1. Enable the required Web services for the cluster: cluster-1::> vserverservices web modify -name spi|ontapi|compat-vserver* -enabled true 2. Create the Web service access service spifor the cluster and assign it the admin role: cluster-1::> vserverservices web access create -name spi-role admin -vservercluster1 3. Create the Web service access service ontapifor the cluster and assign it the admin role: cluster-1::> vserverservices web access create -name ontapi-role admin -vservercluster1 Note: This error will be returned for the admin role, as it is pre-created on cluster create. This error will not be returned if a different role name that is not pre-created is provided. Error: command failed: duplicate entry 4. Create the Web service access service compatfor the cluster and assign it the admin role: cluster-1::> vserverservices web access create -name compat-role admin -vservercluster1 5. Identify the cluster-mgmtLIF for the cluster. This will be used for HTTPS access. A node-mgmtLIF can also be used instead of the cluster-mgmtLIF: cluster-1::*> net intshow -role cluster-mgmt-fields address (network interface show) vserver lif address -------- ----------- ----------- cluster-1 clus_mgmt 10.98.243.120 6. Identify the names of the nodes in the cluster: cluster-1:> node show -fields node node ---------- cluster1-01 cluster1-02 Cluster1-03 Cluster1-04 4 entries were displayed.
6. Add the login to allow http access. cluster1::> login create -username admin -application http -authmethodpassword
cluster1::> login show -username admin -application http (security login show) Authentication Acct Vserver UserName Application Method Role Name Locked ----------- ---------------- ----------- -------------- ---------------- ------ cluster1 admin http password admin no
The following URLs will provide access to the log directory on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/log/ https://10.98.243.120/spi/node-a/etc/log/ https://10.98.243.120/spi/node-b/etc/log/ https://10.98.243.120/spi/node-c/etc/log/ https://10.98.243.120/spi/node-d/etc/log/
The following URLs will provide access to the directory where the core files are saved on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/crash/ https://10.98.243.120/spi/node-a/etc/crash/ https://10.98.243.120/spi/node-b/etc/crash/ https://10.98.243.120/spi/node-c/etc/crash/ https://10.98.243.120/spi/node-d/etc/crash/
On Nov 6, 2013, at 4:54 PM, Scott Miller Scott.Miller@dreamworks.com wrote:
2. RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
As someone else with long NetApp history, but very new to CDOT, I have a related question: syslog.
We have always had a syslog.conf that contained: *.err;kern.* @loghost How does one do the equivalent with CDOT? Is there a shortcut way, or does one need to go manually choose which events to do what with?
Thanks. Davin.
On Thu, Nov 7, 2013 at 9:37 AM, Iluhes iluhes@yahoo.com wrote:
I guess coming from unix background, I always found very useful having ability to have an admin box with all root volumes mounted and just search/grep messages file, copy and edit RC files, rsh scripts via scripts for quota reports, backups, etc. I am sure there might be a way to get some/most/all of that functionality with cDOT, but just the elegance of administering everything from single linux admin hosts just does not seem to be present or fully capable anymore. Am I wrong? I also never worked on anything that need to grow beyond 2-nodes, single cluster. Why would netapp force cluster on such setups, I would have think that majority of install base is small, medium, 2 node clusters
On Wednesday, November 6, 2013 8:08 PM, aprilogi aprilogi@yahoo.com wrote: Just adding a few things:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys) See KB: 1012524
https://kb.netapp.com/support/index?page=content&id=1012542&actp=sea...
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ...” commands
You can also enable remost access so that you can view the logs from http. This is very useful.
See: https://kb.netapp.com/support/index?page=content&id=1013814
Cheers,
—April
To summarize: To enable the web services for a cluster, use these steps from the KB article mentioned above:
- Enable the required Web services for the cluster:
cluster-1::> vserver services web modify -name spi|ontapi|compat -vserver* -enabled true 2. Create the Web service access service spi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name spi -role admin - vserver cluster1 3. Create the Web service access service ontapi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name ontapi -role admin - vserver cluster1 Note: This error will be returned for the admin role, as it is pre-created on cluster create. This error will not be returned if a different role name that is not pre-created is provided. Error: command failed: duplicate entry 4. Create the Web service access service compat for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name compat -role admin - vserver cluster1 5. Identify the cluster-mgmt LIF for the cluster. This will be used for HTTPS access. A node-mgmt LIF can also be used instead of the cluster-mgmtLIF: cluster-1::*> net int show -role cluster-mgmt -fields address (network interface show) vserver lif address -------- ----------- ----------- cluster-1 clus_mgmt 10.98.243.120 6. Identify the names of the nodes in the cluster: cluster-1:> node show -fields node node ---------- cluster1-01 cluster1-02 Cluster1-03 Cluster1-04 4 entries were displayed.
- Add the login to allow http access.
cluster1::> login create -username admin -application http -authmethodpassword
cluster1::> login show -username admin -application http (security login show) Authentication Acct Vserver UserName Application Method Role Name Locked
cluster1 admin http password admin no
The following URLs will provide access to the log directory on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/log/ https://10.98.243.120/spi/node-a/etc/log/ https://10.98.243.120/spi/node-b/etc/log/ https://10.98.243.120/spi/node-c/etc/log/ https://10.98.243.120/spi/node-d/etc/log/
The following URLs will provide access to the directory where the core files are saved on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/crash/ https://10.98.243.120/spi/node-a/etc/crash/ https://10.98.243.120/spi/node-b/etc/crash/ https://10.98.243.120/spi/node-c/etc/crash/ https://10.98.243.120/spi/node-d/etc/crash/
On Nov 6, 2013, at 4:54 PM, Scott Miller Scott.Miller@dreamworks.com wrote:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
Ok - found something and did: event route add-destinations {-severity <=ERROR} -destinations criticals
Davin.
On Thu, Nov 7, 2013 at 4:28 PM, Davin Milun davin.milun@gmail.com wrote:
As someone else with long NetApp history, but very new to CDOT, I have a related question: syslog.
We have always had a syslog.conf that contained: *.err;kern.* @loghost How does one do the equivalent with CDOT? Is there a shortcut way, or does one need to go manually choose which events to do what with?
Thanks. Davin.
On Thu, Nov 7, 2013 at 9:37 AM, Iluhes iluhes@yahoo.com wrote:
I guess coming from unix background, I always found very useful having ability to have an admin box with all root volumes mounted and just search/grep messages file, copy and edit RC files, rsh scripts via scripts for quota reports, backups, etc. I am sure there might be a way to get some/most/all of that functionality with cDOT, but just the elegance of administering everything from single linux admin hosts just does not seem to be present or fully capable anymore. Am I wrong? I also never worked on anything that need to grow beyond 2-nodes, single cluster. Why would netapp force cluster on such setups, I would have think that majority of install base is small, medium, 2 node clusters
On Wednesday, November 6, 2013 8:08 PM, aprilogi aprilogi@yahoo.com wrote: Just adding a few things:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys) See KB: 1012524
https://kb.netapp.com/support/index?page=content&id=1012542&actp=sea...
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ...” commands
You can also enable remost access so that you can view the logs from http. This is very useful.
See: https://kb.netapp.com/support/index?page=content&id=1013814
Cheers,
—April
To summarize: To enable the web services for a cluster, use these steps from the KB article mentioned above:
- Enable the required Web services for the cluster:
cluster-1::> vserver services web modify -name spi|ontapi|compat -vserver* -enabled true 2. Create the Web service access service spi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name spi -role admin - vserver cluster1 3. Create the Web service access service ontapi for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name ontapi -role admin -vserver cluster1 Note: This error will be returned for the admin role, as it is pre-created on cluster create. This error will not be returned if a different role name that is not pre-created is provided. Error: command failed: duplicate entry 4. Create the Web service access service compat for the cluster and assign it the admin role: cluster-1::> vserver services web access create -name compat -role admin -vserver cluster1 5. Identify the cluster-mgmt LIF for the cluster. This will be used for HTTPS access. A node-mgmt LIF can also be used instead of the cluster- mgmt LIF: cluster-1::*> net int show -role cluster-mgmt -fields address (network interface show) vserver lif address -------- ----------- ----------- cluster-1 clus_mgmt 10.98.243.120 6. Identify the names of the nodes in the cluster: cluster-1:> node show -fields node node ---------- cluster1-01 cluster1-02 Cluster1-03 Cluster1-04 4 entries were displayed.
- Add the login to allow http access.
cluster1::> login create -username admin -application http -authmethodpassword
cluster1::> login show -username admin -application http (security login show) Authentication Acct Vserver UserName Application Method Role Name Locked
cluster1 admin http password admin no
The following URLs will provide access to the log directory on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/log/ https://10.98.243.120/spi/node-a/etc/log/ https://10.98.243.120/spi/node-b/etc/log/ https://10.98.243.120/spi/node-c/etc/log/ https://10.98.243.120/spi/node-d/etc/log/
The following URLs will provide access to the directory where the core files are saved on each node: https://<cluster-mgmt-ip>/spi/<node_name>/etc/crash/ https://10.98.243.120/spi/node-a/etc/crash/ https://10.98.243.120/spi/node-b/etc/crash/ https://10.98.243.120/spi/node-c/etc/crash/ https://10.98.243.120/spi/node-d/etc/crash/
On Nov 6, 2013, at 4:54 PM, Scott Miller Scott.Miller@dreamworks.com wrote:
- RSH? Can I run RSH commands?
you can ssh commands, passwordless (once you set up the proper keys)
- No separate exports for qtrees
there is, as of cDOT 8.2.1.
- Not clear how to access log file of the cluster....mounting root
share not possible, right? only for vServer..
"event log ..." commands
Toasters mailing list Toasters@teaparty.net http://www.teaparty.net/mailman/listinfo/toasters
-
April Jenner -
aprilogi
-
Davin Milun -
Iluhes -
James Andrewartha -
Jeremy Webber -
John Clear -
John Stoffel -
Koji Kabuto -
Martin -
Scott Miller -
Sebastian Goetze -
steve klise -
tmac