Tips and Tweaks
Do you have older storage on your floor taking up space and is too costly to replace? Why not use a N series Gateway to help add value to that asset and ‘recycle’ the storage for other uses. We are all looking for ways to do more with fewer resources and demand on IT keeps getting larger.
N series storage can use older disk subsystems at the same time using its own native disks. It allows clients to use technology like deduplication, snapshots and thin provisioning with out having to rip and replace your existing footprint. The Gateway uses the underlying storage structure’s RAID for protection so there is no bloat with additional raid from the gateway itself.
If you have multiple storage units, no problem. We can connect via Fibre Channel to multiple storage platforms. This allows us to create teirs of storage for your applications so that slower systems are used for archives and faster systems are used for production data. Once the gateway is configured, the N series will present data just as if the disk were native.
We also add value for systems that don’t support certain technologies. For example, if you are interested in moving into 10Gbps networking. N series is the only platform that supports it at this time. You could use your storage system that you have now, put a gateway in front of it with a 10Gbps card and voila’. It is that easy.
Now you don’t have to use older storage appliance behind a gateway, we are using gateways in front of XiV, SVC and DS5k /8k. Now if you are looking for just Fibre Channel attachment this might not be a good fit but if you are looking for multiprotocol, application integration with Exchange/SQL/Oracle/VMware and the like. We can use the underlying storage system and make it more productive with the N series portfolio.
Other uses for a Gateway include disaster recovery, archiving, VDI projects, Dev and Test environments and much more. We see Gateways being used for data migrations, for site to site mirroring and recovery of data centers.
For more information about N series Gateways, check out the IBM site for N series
Posted by richswain in Oct 13, 2009, under Tips and Tweaks
A packet is sent down the line and it comes to a fork in the channel, does it pick the right way? Only if you tell it.
With N series active/active configurations, you can see a lun from both controllers. This sometimes confusses people and they end up picking the wrong path to the hosting controller. The I/O that should be going directly to the primary controller is now going through the ‘partner’ and over the interconnect. This does increase the amount of workload each system has to put out as it handles the extra traffic.
Typically there are a few things to check when trying to diagnose the issue.
- Identify the LUNs being accessed through the partner node’s FCP Target port.
a. lun stats -o (LUN STATISTICS)
- Identify the host initiators that are performing the I/O through the partner path
a. lun config_check -A (LUN CONFIG CHECK)
b. lun show -v (LUN CONFIGURATION)
c. igroup show -v (INITIATOR GROUPS)
- Identify the primary storage controller FCP target ports available for access to the LUN.
a. fcp show cfmode (FCP CFMODE)
b. fcp show adapters (FCP TARGET ADAPTERS)
- Verify the host initiator connectivity to primary FCP target ports and the host MPIO software configuration.
- Verify use of the partner path has ceased from both cluster nodes.
a. sysstat -b 1
Also make sure your MPIO settings are correct and not affecting the system path. All hosts should also be running the host utility kit provided by IBM.
With these techniques you should be able to clear up any mis-directed I/O. As with any correction, if you are not sure please contact IBM Support at 1.800.IBM.SERV or open a service request ticket at IBM Online Support Portal
Posted by richswain in Oct 12, 2009, under Tips and Tweaks
You just received your new N series system and after plugging it all up, you want to read up on how configure autosupport. Or maybe, you want to chat with others about the best way to setup a cifs share. Whether you are an experienced user or just starting out, we have all of the information you need to get through the install, administrate the day to day activities and if needed a recovery.
I put together a list of topics / links people are always asking me to send to them.
Your N series system has to be registered with the current information so IBM Support can contact you and you can get access the IBM Support Page. To begin this process go to the following link and start your process with the overview of this process in step 1.
You can now customize your support page with IBM Support Portal. Here you can select your different IBM products (storage, servers, software…).
My IBM is where you can sign up for notifications on IBM products.
To download code from IBM you have to be signed into the support site and choose Data ONAP as your storage family, then click download.
On the next screen you will click on the link at the bottom for Software Packages for registered users.
The following pages allow you to choose your product to download. If you do not see a product that you purchased on the page, contact your local account team or call 1.800. IBM.SERV for support (See below for more instructions on contacting support via phone).
All of the IBM documentation is based on the release of Data Ontap. The best way to get to the documentation is the through the ‘Matrix’. From the same screen you used to get the software download, click on the N series Data ONTAP Matrix link at the bottom:
From here you find release notes and publications that will assist you with every aspect of your N series system and environment.
Here is a quick link to the latest (October 12, 2009) release of publications. You will have to either be logged in or the site will prompt you to login.
This page has tons of information that you will need for administrating and recovering your system and more. Here is a list of topics:
- General Information
- Data Access Protocols
- Storage Management
- Data Protection/Recovery
- Command References
- Autosupport and Troubleshooting
Ask questions and chat with IBM team members about N series and other products:
Professionally written Redbooks and Redpapers on N series and other technologies.
A couple of suggestions:
There are over 60 Redbooks written about N series and its software components. Need help with virtualization, Exchange, SQL, SAP? This is where you will find it.
Opening a service ticket:
Currently there are two ways to open a service ticket with IBM support; calling 1.800.IBM.SERV or through the website.
1. Calling IBM.SERV
When calling IBM.SERV you will have two options: Hardware or Software. Your warranty for break fix covers both the system and the software that was sold to you. If you are needing a part fixed or if an error is occurring with your software, you will choose option 1 for Hardware.
If you have purchased a supportline contract from your account manager, then you will be able to ask IBM any type of installation or configuration question by choosing option 2 for Software.
Typically if you do not know which option to choose, then select option 1 and our Level 1 team will transfer you to the appropriate team as needed.
As with any IBM production you will be asked for the Machine type and serial when calling. This can be found on the front bezel of the machine or in the header of your autosupport:
===== SYSCONFIG-A =====
> Data ONTAP Release 126.96.36.199: Wed Jun 25 11:01:02 PDT 2008 (IBM)
> System ID: 0101224997 (toasterA); partner ID: 01014303234 (toasterB)
> System Serial Number: 2864130990522 (toasterA)
Look for the Serial Number listing, the first four digits are your model, the next 7 digists are the serial number.
2. Through Webpage
If you wish to submit a ticket to through the IBM Support site simply click on the link in the blue box (on the right side of the page).
After answering your questions the ticket will be submitted to the IBM Support team who will call you to resolve your issue.
If for any reason you do not receive the level of support you deserve, you can always call the 1.800.IBM.SERV number and ask to speak to Duty Manger. This person will always be available and can be an advocate for you when you need someone.
Pick up any industry magazine or your favorite blog and you are sure to find them talking about how FCoE and FC over Convergence Enhanced Ethernet. We have been consolidating the servers into virtual machines and consolidating our storage at the same time, but now we can do the same with our networks.
We have 10 GB networks coming into the market place and they are gaining in popularity as network admins replace older technology with newer, just like when 1 GBPS switches came out to replace our 100 mbps one. But 10 GBPS actually goes beyond the speed of the network. We can now use fewer ports and fewer switches, which down the road leads to lower costs for power, cooling and rack space.
N series supports FCoE as well as iSCSI, NFS, CIFS all over the 10GBPS link. We also support VLAN tagging to help separate the different protocol workloads. Here is a great Redpaper that will explain FCoE and how it can help you.
There have been plenty of you that have said the IBM support site is hard to navigate. You wanted a quick way of finding things that only pertain to your environment. Finally, IBM now has a way to do both, introducing the IBM Support Portal.
The portal is very user friendly and allows me to pick up to t10 different items that I want to keep up to date on. For Example, If I have a N7900 and a HS20 Blade and some System X server, I can choose all three on my portal and get information on the latest updates, news and flashes.
When you click on the link it will take you through a couple of simple questions to get an idea of what things you are looking for on the IBM support site. Once this is done you can then customize the portal with different types of downloads, forums, tools and so much more.
A well written blog was posted today on the Netapp site about best practices with Hyper-V. Chaffie McKenna is an architect for Microsoft Solutions Engineering. The article goes into different parts of setting up the network and the vms.
If you are looking to deploy Hyper-V in your environment you need to read this article.
Posted by richswain in Sep 28, 2009, under Tips and Tweaks
Now that VMWorld ’09 is over, there are tons of people talking, blogging, tweeting about everything VM. If you are looking at upgrading to vSphere, or just starting out there are some things you will want to know how to configure. As with any storage, there are different ways to install and configure and N series is no different.
I found this great Best Practice document from Netapp that goes through everything from how to setup the swap space to multipathing. If you are looking at going to the new VM platform, you need to read this document and take notes.
Posted by richswain in Sep 23, 2009, under Tips and Tweaks
What are the signs of a good recovery or business continuance plan? If you have ever had to develop a DR plan you know you to not make any assumptions and everything is laid out in perfect order on the road to recovery. But those who have had to implement a DR plan know the road to recovery is paved with great intentions. We can not predict everything that could happen and the recovery points for those systems.
When I talk to IT admins about their DR plan, most say they have one, some say they have tested it. Those who did try out their DR plans had to make so many changes to the environment, they had questions if it was a good test or not. While doing some research on this topic I came across an interesting fact: 6 percent of companies suffering from a catastrophic data loss survive, while 43 percent never reopen and 51 percent close within two years. In large companies, the recovery budget usually runs between 2 percent and 4 percent of total IS costs.
With so few dollars being spent on the recovery and the risks of actually surviving a data loss, we have to find solutions that are sound but cost effective. With flex-clone technology from IBM N series, you can mount a snapshot of your data as a clone of the active file system. This could be a virtual machine to an exchange database. A clone does not consume any space unless the data is changed or new data is added. For example, a test of the SQL database that is 20GB, once the clone is created the clone consumes 0 GB amount of space. If the cloned database is attached to the server and your DB does some testing by adding some new records, say 10mb, the total amount of space needed to test the recovery is only 10mb.
Storage administrators can perform this test on the primary system or a secondary system with our snapmirror option. Snapmirror allows an admin to asynchronous or synchronously send data to another N series unit. Typically the secondary site is smaller system with mainly cheaper and denser drives. Mirrored data sits on the secondary system as read only but can be cloned and tested with out interrupting the mirror relationship. The setup is simple and has some throttle mechanisms to not saturate your connection between sites.
For those of you who run VMWare, you have all of this scripted and tested with Site Recovery Manager. SRM will use both flex clone and snapmirror technology to allow you to test or simulate a failure in the primary site. More information can be found on youtube.
If you have a N series system now, all of this technology is built in and we can turn it on using a key. If you want to ‘kick the tires’ and see how this technology will help you make your DR plan a solid realistic plan, contact your local IBM Storage Sales rep.
Posted by richswain in Sep 17, 2009, under Tips and Tweaks
Consider making a reservation a nice restaurant. You and your friends decide to go out to dinner this coming Friday at one of the newer restaurants. If you just show up you know there will be a line, so you call them ahead of time and ask them to hold a table of 6 at 7:00pm, this coming Friday. When Friday comes and your entourage shows up at the restaurant you are ushered in and sat immediately with no waiting.
What would happen if only three of the six people show up, and the restaurant sits you at table for six people. That restaurant has now lost those three places and will not be able to re-use them until your party leaves. Also, the party of 4, 5 or 6 that is waiting for a table now has to wait until another table opens up.
We can draw similarities between our restaurant and our N series storage. When clients leave the fractional reserve set to 100% for their luns, we have to hold the same amount of space as the size of the lun. If the lun is 100GB and fractional reserve is set to 100%, we will save 100GB in reserve to accommodate it.
The reason for reserving space is simple, it’s a guarantee that writes to the lun will always succeed without deleting snapshot copies. With a 100% fractional reserve, a lun can be completely re-written and you will not have to delete any snapshots. By lowering the percentage, you lower the amount of space guarantee for an overwrite.
For example: A 100GB lun with a ROC of 10GB per day
You take one snap per day and want to keep 10 days of snapshots
Space for Snapshots = 10GB ROC x 10 days of snapsshots = 100GB
You want to continue to write to the Lun through the weekend, even after the last snapshot and there is no free space.
Overwrite Space: 10GB ROC x 2 DAYS = 20 GB
Volume size = 100GB lun + 100GB snapshots + 20GB reserve = 220GB
Fractional Reserve: 20GB / 100GB = 20%
To see what your fractional reserve is set to do the following:
Nseries> vol options vol0
root, diskroot, nosnap=on, nosnapdir=on, minra=off,
no_atime_update=off, nvfail=off, ignore_inconsistent=off,
snapmirrored=off, create_ucode=off, convert_ucode=off,
maxdirsize=31457, schedsnapname=ordinal, fs_size_fixed=off,
compression=off, guarantee=volume, svo_enable=off, svo_checksum=off,
svo_allow_rman=off, svo_reject_errors=off, no_i2p=off,
fractional_reserve=100, extent=off, try_first=volume_grow,
To change the value simply type the following command:
Nseries> vol options vol0 fractional_reserve 20
If you are uncomfortable with changing this value, all of this information is in the “Block Access Management Guide“. Also, feel free to call your local IBM technician or 1.800.IBM.SERV or leave a comment here!
Posted by richswain in Sep 15, 2009, under Tips and Tweaks
There is a very good NEW Demo for you to share. “Honey, I shrunk the LUN”. What this demo shows is SnapDrive 6.2, which is now in Beta, shrinking a LUN on the fly without causing so much as a blip on the OS.
Every single LUN in existence is over provisioned. Sometimes by 100%, or more. It has had to be this way. It is, by far, the biggest reason for wasted space on a disk array. And it is worse on traditional storage. When migrated to N series the problem will still be there, until now.
N series provides customers the ability to win back storage held hostage by traditional array technologies and the provisioning it forces upon them – without downtime or disruption.
A customer with 50 TB of storage today, may be able to buy just 25 TB of N series storage (or less), because we can shrink their LUNS and gain back that which was written off as lost.
A customer may be able to leverage Dedupe and gain back even more.
You could use this against traditional storage as part of a gateway campaign, and help them extend the life of their traditional storage array
With N series they can begin with Thin Provisioning, or they can simply switch it on at a later time. No migrations required. No disruption.
Thin Provisioning, Dedupe and LUN Shrink all work together. On all N series systems. At the same time.
Here is the link for the YouTube Demo:
Let me know what you think!