Developers Club geek daily blog

Use of LRDIMM modules in high-end servers

1 year, 1 month ago
LRDIMM (Load-Reduced Dual Inline Memory Module or "DIMM with reduced loading") – type of the modules of memory supported by server platforms since 2012. LRDIMM modules are similar to register DIMM modules and approach the same connectors of memory. However the principle of work of LRDIMM differs from RDIMM. Using LRDIMM in the normal server it is possible to make 512 GB, 1tb or 1,5tb memories.

Read more »

FlexPod DataCenter: Direct-Attached Storage

1 year, 1 month ago
In the previous article I told about "a не-FlexPod of DC" to architecture which can be supported from "one hands" according to the Cisco "Solution Support for Critical Infrastructure" (SSCI) program. Its main feature consists that in it there are no Nexus series switches and if there to add them, such architecture can become full-fledged FlexPod DataCenter.

Here the speech about new design of a network, for FlexPod DataCenter, with a live broadcast of SHD NetApp will go to the UCS domain. Difference from standard architecture of FlexPod DataCenter is that the switches Nexus are located not between UCS and NetApp, and "over" UCS.

In spite of the fact that and before SHD NetApp of the FAS series it was possible to connect directly to Fabric Interconnect (FI), officially the architecture of FlexPod DataCenter did not predusmarivat such design. Now the design with a live broadcast is supported and to saportitsya as FlexPod DataCenter architecture.

The general design of the FC and FCoE network with a live broadcast
Opisny switching circuits on the image are higher
Simultaneous connection on FC and FCoE is represented for two reasons:
  1. It is so really possible to make and it will work
  2. To show that it is possible for FC and/or FCoE.

Ethernet connection between two NetApp FAS controllers are represented for two reasons:
  1. To show that it is two notes of one NA of system (if it is more notes, on the picture surely there would be cluster switches).
  2. External cluster link obligatory accessory of the Clustered DataONTAP operating system.

FC a link from FI to Nexus of the switch is represented for two reasons:
  1. For the future. When we need to switch NetApp to Nexus switches and FI got access to the Lun'am. Then the scheme will become more scaled, it will be possible to add still UCS domains.
  2. To take away resources from storage from other servers which do not enter the UCS domain. For example UCS Rack of servers (UCS C a series) not connected to FI or servers of other vendors.

For traffic Ethernet jointly as with a live broadcast and the iSCSI protocol, and a live broadcast and the FCP protocol — by means of the multipasing which is built in these protocols there are no problems in setup of fault tolerance and balancing on links.
And here for NAS protocols, with a live broadcast (NFS v2/NFS v3 and CIFS v1/CIFS v2), in a type of lack of balancing of loading and a multipasing in these protocols, their function have to fulfill some other, underlying protocols, such as LACP and vPC (FI does not support vPC), thus fault tolerance for Ethernet of a network will have to build somehow in a different way. For example fault tolerance for Ethernet can be made at the level of a virtual switch (that can have problems with performance of such switch) or by means of active-passive of switching of the aggregated network link, without LACP (that the traffic on all available links will not allow to balance), ifgrp link for this purpose aggregated, from SHD, has to be configured in the single-mode mode.
The question with a live broadcast for NAS protocols not so sharply looks for NFS v4 and CIFS v3.0, but demands support of these protocols on the party of clients and SHD (all systems of FAS with cDOT are supported by NFS v4 and CIFS v3.0) as both protocols at last purchased a certain similarity of a multipasing.
to configure FCoE and CIFS/NFS a traffic over one link
  • First the version of a firmware Cisco UCS firmware 2.1 or above is necessary
  • Secondly the storage with 10GB CNA/UTA ports is necessary

Further we go on settings:
From NetApp storage it is necessary to transfer ports to CNA status (existence of CNA ports, normal Ethernet 1/10Gbs is necessary ports of it do not support), by means of the ucadmin command on SHD (restart of SHD will be required). In system independently "virtual" ports Ethernet and "virtual" FC ports, separately will be displayed (though the physical port for one such "virtual" Ethernet and one "virtual" FC will be used one). Such ports separately as normal physical ports are configured.
On FI it is necessary to include the FC mode in a status of "Switching mode", in the Fabric A/B settings on the Equipment tab. This setup will demand restart of FI.
After restart of FI on the Equipment tab it will be necessary to transfer convergent ports to the Appliance port mode, after several seconds the port will pass into online mode. Then recustomize port in the FCoE Storage Port mode, on the right panel you will see type of Unified Storage port. Now will be vozmozhnocht to select VSAN and VLAN for such port. And the important point created earlier VSAN has to have included "FC zoning" on FI to execute a zoning.

Setup of a zoning for FI:
SAN-> Storage Cloud-> Fabric X-> VSANs-> Create "NetApp-VSAN-600"->
VSAN ID: 600
FCoE VLAN ID: 3402
FC Zonning Settings: FC Zonning-> Enabled

SAN-> Policies-> vHBA Templates-> Create "vHBA-T1"-> VSAN "NetApp-VSAN-600"

SAN-> Policies-> Storage Connectivity Policies-> Create "My-NetApp-Connectivity"-> Zoning Type-> Sist (or Simt if is necessary)-> Create->
FC Target Endpoint: "NetApp LIF's WWPN" (begins with 20:)

SAN-> Policies-> SAN Connectivity Policies-> Create "NetApp-Connectivity-Pol1"-> vHBA Initiator Group->
Create "iGroup1"-> Select vHBA Initiators "vHBA-T1"
Select Storage Connectivity Policy: "My-NetApp-Connectivity"

During creation of Server Profile to use the created politicians and vHBA a template.

Read more »

Data record in Teradata with the help of DML

1 year, 1 month ago
If to go straight all yes directly, far you will not leave …
(Little prince, Antoine de Saint-Exupéry)

Recently the colleague somehow appealed to me to help with Teradata. It is actively implemented now and the first step of this implementation is data loading on a daily basis. It is necessary to fill in much and as soon as possible. Some alternative methods of data loading asked to find me in Teradata which not strongly would depend on the resources selected to the user. In the course of this work I had to become more closely acquainted with .NET Data Provider for Teradata. An input of acquaintance some curious parts which knowledge, in my opinion, can be very useful became clear. As not all know about Teradata, I will begin with its short description.

Read more »

NetApp becomes even more effective with inline dedupe

1 year, 1 month ago
The Data ONTAP 8.3.2RC1 version became a few days ago available. "RC" means Release Candidate and consequently, according to the rules of naming of versions accepted in NetApp, this release already passed all internal tests and can be used by customers not only for an assessment of new opportunities, but also for a produktiv, including systems at which business and critical tasks work. Support of vendor completely extends to the systems working at "RC" Data ONTAP versions.

Sellers of NetApp (yes, it and we too ;) are not tired to remind (and do it perfectly) that high utilization of disk resources in the NetApp systems is reached due to use of software of optimization — a deduplication and a compression.


Read more »

Example of express performance review of SHD by means of the free Mitrend service

1 year, 1 month ago

Studying of a problem with performance and search of solutions are familiar to much not by hearsay. There is a large number of instruments of visualization and parsing of statistics of input-output. Now automation of the intellectual analysis based on Internet services gains steam.

In this post I want to share an example of the analysis of a problem with SHD performance based on one of such services (Mitrend) and I will offer ways of its solution. In my opinion, this example represents the interesting etude which as I think, can be useful to a wide range of IT readers.

So, the customer asked EMC to look at performance unrolled at it in SAN of hybrid storage system VNX5500. The VMware servers on which all turns "in general" are connected to SHD: from infrastructure tasks to file a sphere and the DB servers. Complaints to podvisaniye of the applications unrolled on connected to VNX servers were the cause of carrying out this express assessment.

Read more »

You give a flash revolution! As quickly and safely to pass on flash drives

1 year, 1 month ago
Hi, Habr! Today we would like to tell about a new solution for acceleration of transition to the data processing centers using only flash drives. The effective HPE 3PAR StoreServ technology along with protection of StoreOnce will help clients with the innovation upgrade of data-centers. But give about everything one after another.

Read more »

Future of data storage: 8 new technologies

1 year, 1 month ago

In our blog on Habré we not only tell a selection of high-performance and capacious storage devices about development of the cloud service, and also told about a 16-terabyte disk from Samsung and new approach to creation a flash memory from Intel.

The industry of data storage endures a stage of drama changes which will be shown as soon, and in the long term. Therefore today we will talk about technologies which will lead to these changes.

Read more »

New Samsung of 950 Pro PCIe SSD: everything that should be known

1 year, 1 month ago
The Samsung Electronics company declared start of the first solid-state drive of Samsung of a series 950 PRO of a form factor of M.2 with the scalable NVM Express controller (Non – Volatile Memory Express). Let's understand why Samsung passes to a new form factor and the controller, and also than it is good for the user.

Read more »

Why to be updated to Data ONTAP Cluster Mode?

1 year, 1 month ago
As I already wrote in the previous posts, Data ONTAP v8.3.x is one of the most significant releases of an operating system for FAS series NetApp storage systems.

In this article I will give the most significant, from my point of view, new functions of storage systems NetApp in most current release of Clustered Data ONTAP. By tradition I will give an example on cars: Provide you have a Tesla the car, you updated a firmware and received an auto pilot with an autoparking free of charge though it was not there earlier. Truth pleasantly? And so the most important arguments to update your system to Cluster-Mode is saving of investments and an opportunity to receive the most modern functionality on old iron:

  • Online detection (deduplication) of zero on the run that can be very useful in case of a DB and provisioning of virtual computers.
  • Online a deduplication for FlashPool (and AFF) systems that will allow to prolong service life of SSD disks. Function is available since 8.3.2.
  • If to be updated to VMWare vSphere 6, you will have a support of vVOL as with NAS and SAN
  • Support of NFS4.1 which is also present at VMware vSphere 6
  • Support of pNFS which allows to parallelize NFS and to switch between ways from the client to file a sphere without its reassembling, is supported with RHEL 6.4 above.
  • Support of SMB (CIFS) 3.0 which works with clients since Win 8 and Win 2012
  • Support of closing of files and sessions for SMB 3.0 from Data ONTAP
  • Support of SMB of 3.0 Encription.
  • SMB Continuous Availability (SMB CA), gives a switching opportunity between ways and controllers of storage without rupture of connection that is very important for work of SQL/Hyper-V
  • ODX during the work with Microsoft SAN/NAS allows to unload routine tasks, type to hammer a data unit with a certain pattern, and allows not to drive excess data between a host and storage.
  • Online migration of volyyum on units, including on other notes of a cluster
  • Online migration of lun after volyyuma, including on other notes of a cluster
  • Online switching of units between notes HA vapors
  • An opportunity to integrate heterogeneous systems in one cluster. Thus the upgrade is performed without stop of data access, thanks to such possibility of NetApp calls the cluster Immortal. At the time of updating of a cluster, its note can consist of different cDOT versions. I cannot miss an opportunity and not mention that most of competitors if in general has a clustering, then it first is very limited on number of notes, and in-torykh all notes of a cluster are obliged to be identical (a homogeneous cluster).
  • ADP StoragePool — technology for more rational distribution of SSD under a cache (hybrid units). For example at you is only 4 SSD, and you want that 2, 3 or four units got advantage from a caching on SSD.
  • ADP Root-Data Partitioning will allow to refuse the selected root of units for the FAS22XX/25XX and AFF8XXX systems
  • Space Reclamation for SAN — returns remote units to storage. I will remind that without SCSI3 UNMAP to a dezha if on yours to the moon data units were removed, on thin to the moon in the storage these blocks were all the same marked as used and occupied disk space, and any thin the moon could grow earlier only as earlier just there was no feedback mechanism of storage and a host. For support of Space Reclamation hosts have to be ESXi 5.1 or above, Win 2012 or above, RHEL 6.2 or above.
  • Adaptive compression — improves the speed of reading the compressed data.
  • Improvements of work of FlexClone for files and lun. There was a possibility of a task the politician of removal of clones of files or lun (it will be useful for example with vVOL).
  • An opportunity to authenticate the SHD administrators by means of the Active Directory (the license CIFS is not required).
  • Support of Kerberos 5: 128-bit AES and 256-bit AES enciphering, support of IPv6.
  • Support of SVM DR (on the basis of SnapMirror). I.e. an opportunity to otreplitsirovat all SVM on the reserve website. An important point is an opportunity at a stage of setup of the relations of replication to preset new network addresses (Identity discard mode) as on a reserve site, ranges of network addresses, other than the main site, are often used. The Identity discard function will be very convenient not to the big companies which are not able to afford the equipment and communication channels to stretch L2 the domain from the main site on spare. In order that clients switched to new network addresses enough to change the records DNS (that can be easy avmomatizirovano at momoshch of a simple script). Also Identity preserve the mode when all settings LIF, volume, LUN remain on a remote site is supported.
  • Possibility of recovery of the file or the moon from a backup copy of SnapVault without recovering all volyyum.
  • An opportunity to integrate SHD with anti-virus systems for check file a sphere. Computer Associates, McAfee, Sophos, Symantec, Trend Micro and Kaspersky are supported.
  • Work of FlashPool/FlashCache is optimized. Allows to cache the compressed data and big blocks (earlier both of these data types did not get to a cache).

Read more »

HV data storage format as attempt of a solution of the problem of visual storage of text boxes

1 year, 1 month ago

Not so long ago before me there was a task to have an opportunity to store data in text form that not only the program worked with them, but the person could read and edit (and also to create from scratch in a text editor). For this purpose there is already a set of convenient and good formats, for example JSON, YAML, XML and so on. But in the considered systems the moments which, nevertheless, were not pleasant a little came across.

I will pay special attention to bright inconvenience of the majority of such formats (naturally, in my opinion) including very powerful and popular — the problem connected with storage of the text: how to write a text box which may contain any text characters that its contents had not to be changed and it did not influence parsing, also the different substrings matching office combinations and different non-standard indents can meet there.

Read more »