Archive for rackmounts
May 11, 2008 at 7:49 pm · Filed under blades, rackmounts, virtualization, vmware
http://communities.vmware.com/message/940884?tstart=0#940884
Check out this great post on the VMWare community talking about blades or rackmounts for virtualization. I’ve got mixed opinions on this. Rack servers can be a great vehicle for virtualization and I have used them in the past. This question is going to be dependent on your exposure and ‘comfort’ with blades as a platform and also volume. I see blades as more suited to virtualization projects when we’re using the volume farm type scenario. Putting in a blade solution with the integrated switches, with the right processor and memory configuration could be a powerful solution, it’s going to be dependent on the range of networks involved. You will find that blades require economies of scale, by that I mean (with no negativity) a blade solution is going to be cheaper the more blades you fit in the enclosure, because of the cost of the enclosure.
The conversation amongst the vendors and the end users are set to continue, I see blades more a volume virtualization tool. The example I like to use is one of the enterprises organizations that I spoke to last year, they were taking 800 rack servers ranging from Compaq 2500s to current day HP Proliants and virtualizing them on to blade technologies for the following reasons:
Blades allowed them to have one hardware platform - let’s say they decided on 400 IBM blades, they changed their hardware support contract and reduced it substantially, instead of 16 derivatives of servers to support there were two, the IBM blade and the enclosure.
The blades were configured for SAN boot, so a blade could be re-provisioned or the storage re-mapped in the event of a hardware failure.
The debate is set to continue, the key is to focus on what your goals are, where your strengths are and go in that direction. I’m all for blade technology as a platform in grid, hpc or as an alternative to a rack server, but they need to be used (as with anything) to the benefit of the customer first, IT second.
May 11, 2008 at 8:47 am · Filed under HP BL30P blade, blades, rackmounts
http://web.mac.com/martinmacleod/iWeb/Site/Blog/ilomanagement.pdf
I’ve written a quick ILO user guide. It’s simply an overview of the options and shows how you upgrade the firmware. I hope I haven’t missed anything or made any mistakes. In the meantime click the above url to view/download it and feel free to comment.
May 8, 2008 at 4:40 pm · Filed under rackmounts
http://www.finextra.com/fullstory.asp?id=18441
Banking giant HSBC says a computer server holding the confidential personal details of around 159,000 customers has gone missing from a branch in Hong Kong.
The server - which contains account numbers, customer names and transaction details - disappeared on 26 April during renovation work at a Kwun Tong branch in east Kowloon on April 2.
In a statement, the bank says the server does not contain any customer PINs, passwords or user IDs and has “multiple layers of security”.
Vincent Cheng, chairman, China and Asia operations, HSBC, told reporters that affected customers would not be liable for any losses arising from the situation.
Check out this article, the need to protect your customer data and at the same time reduce your operational risk, is not just part of good business practice, of good systems administration, but increasingly an important aspect from the view of the customer.
Data loss can happen to everyone, what processes you have in place, and how they are enforced to limit any potential liability or loss is what matters.
April 26, 2008 at 11:35 am · Filed under blades, rackmounts
http://e4services.homestead.com/files/Support/ProliantHardDriveInstallReplacement.html
Each SCSI channel on the controller supports up to 14 drives. Drives can be of the Wide Ultra3 or Wide Ultra2 type.
Each drive on a SCSI bus must have a unique ID value in the range 0 to 15 (except ID 7, which is reserved for controller use). This value is set automatically on hot-pluggable drives in ProLiant servers and storage systems, but values for other drives must be set manually.
- Do not terminate the drives. HP and Compaq branded servers and internal cabling provide the required termination of the SCSI bus.
- Do not use drives of different capacity in the same array. The excess capacity of larger drives cannot be used by the array and is wasted.
- Do not use hot-pluggable drives on the same SCSI bus as non-hot-pluggable drives.
Always be careful when swapping or moving drives about. Using RAID as a backup procedure is something to think about, but do ensure you have an adequate backup, pressing F1 when you should have pressed F2 could be the difference between rebuilding the server and quickly rolling back those changes. This article goes over some of the topics in replacing hard drives in a Compaq/HP Proliant - if in doubt speak with your support team/vendor before proceeding and ensure the firmware is up to date.
April 21, 2008 at 10:13 pm · Filed under blades, rackmounts
http://www.dailymirror.lk/DM_BLOG/Sections/frmNewsDetailView.aspx?ARTID=12447
HP recently announced that in the past 18 months, more than 50 Asia Pacific and Japan customers from have successfully migrated from legacy mainframe systems to HP Integrity systems to support mission critical IT environments. Companies moving to next-generation data centers from diverse industry sectors have replaced legacy systems, including IBM mainframes, with HP Integrity to minimize high hardware maintenance and software licensing costs.
Fueled by increased demand for HP Integrity and Integrity NonStop servers in Q4 2007, HP’s Intel Itanium (EPIC) processor-based server revenues grew 53.0 percent in Asia Pacific (excl. Japan) region(1). According to IDC, HP tied for the top position in the Asia/Pacific (inc. Japan) Unix server market in revenue terms, securing 33.0 percent market share measured in CYQ4, 2007. HP’s RISC & EPIC Unix server revenues grew 29.3 percent on a quarter-on-quarter basis in Q4 2007, underpinned by 23.8 percent growth in EPIC Unix. RISC Unix server revenues grew by 46.4 percent on a quarter-on-quarter basis, outperforming the market growth of 8 percent during the same period. HP Integrity servers represented the majority of HP’s Business Critical Systems revenue in Asia Pacific (Inc. Japan), with 77.0 percent share in Q4 2007 (2).
Very cool, this article illustrates the sales of Itanium based servers and also highlights the different solutions which have been deployed using Itanium based servers, an interesting read, check it out. Itanium might not be for everyone, but deployed in the right way (as with anything else) with the right layered applications/middle ware, they can be an effective solution.
April 21, 2008 at 10:03 pm · Filed under rackmounts
http://www.informationweek.com/news/hardware/processors/showArticle.jhtml?articleID=207400342&subSection=News
Fujitsu on Thursday introduced a line of high-performance servers powered by Intel (NSDQ: INTC)’s latest Itanium dual-core processor.
The Primequest 520A, 540A, and 580A servers are available with up to eight, 16, and 32 9100 series processors, respectively. The product line is expected to ship in May.
The servers are available with either open source Linux or Microsoft (NSDQ: MSFT)’s Windows Server. The systems are built for running databases, enterprise resource planning applications, and scientific computing.
The new products are updates of the midrange Primequest 520 model and the high-end 540 and 580 models. The three systems were introduced in July 2006.
Very cool, the more cross platform solutions we have for the end user, the more choice we have in terms of the processor and platform combinations which has to be a good thing for competition in the market and for the end user in finding the solution for them. That Fujitsu have updated their Itanium servers is very cool, I wonder if there will be a Fujitsu Itanium blade version as well some time soon?
April 14, 2008 at 11:46 pm · Filed under rackmounts
http://www.prweb.com/releases/enterprise/sparc/prweb844994.htm
SANTA CLARA, Calif. & TOKYO (Business Wire EON/PRWEB ) April 9, 2008 — Sun Microsystems, Inc. (NASDAQ:JAVA) and Fujitsu Limited (TOKYO:6702) today expanded the Sun and Fujitsu SPARC(R) Enterprise server line with the introduction of two new systems based on the UltraSPARC(R) T2 Plus processor. Scaling from the edge of the network to the heart of the enterprise, the third-generation CMT SPARC Enterprise T5140 and T5240 servers deliver breakthrough performance and scalability and help enable customers to consolidate the datacenter into an ultra-dense, energy efficient compute environment, optimized and managed by the Solaris(TM)10 Operating System (OS). The SPARC Enterprise T5140 and T5240 servers deliver up to 16 times higher compute density than competitive two-socket x86 systems and up to 32 times higher compute density than competitive four-socket x86 systems. The Sun and Fujitsu SPARC Enterprise T5140 and T5240 servers provide up to three times more performance than competitive RISC systems in half the space.
This is great news, these new servers bring more performance to the plarform, and more choice to the consumer which has to be a good thing. Do check out the article for more information, it’s got some interesting comments about performance per watt and example application performance, it’s always interesting to read these kind of things.
April 14, 2008 at 11:36 pm · Filed under blades, rackmounts
http://www.midmarket.eweek.com/c/a/News/HP-Offers-LowCost-Servers/
Hewlett-Packard is offering a new low-cost rack-mount server and blade as well as new CRM tools.
Hewlett-Packard, which has been building up a line-up of inexpensive hardware and software directed at businesses with limited IT resources, is now launching a pair of low-cost servers to add to the portfolio.
The x86-based hardware includes a new blade server, the ProLiant BL260c G5, and the ProLiant DL120—a 1U (1.75-inch), single-socket rack-mount system. Both systems offer a number of low-cost configurations, including a choice of Intel processors that range from the newer Xeon 5200 and 5400 series microprocessors down to the older Celeron and Pentium chips.
Very cool, these new low-cost configuration servers might be ideal for a range of solutions or customers. Whether it’s a server for a specific use (DC/mail server for a small business), or a low-cost unit for that development blade farm, the low-cost option has to be a good thing for the end user. I’ll need to check them out. The DL120G5 quick specs are here, the BL260c G5 quick specs are here.
April 10, 2008 at 11:01 pm · Filed under How IT works, rackmounts
http://www.pressebox.de/pressemeldungen/ibm-deutschland-gmbh-4/boxid-165608.html
(pressebox) SAN FRANCISCO, 09.04.2008 - At a customer event here today, IBM (NYSE:IBM) announced two high-end Power Systems models - the world’s fastest UNIX® server and a unique water-cooled supercomputer. The new systems offer sophisticated IBM virtualization technology and energy-saving capabilities to help dramatically reduce bottom-line operating costs, such as those for energy, floor space and systems management, while improving system performance, helping customers transition to a new enterprise data center.
The new UNIX enterprise server, the Power(TM) 595, designed to extend IBM’s leadership in the UNIX market, will be attractive to existing IBM clients as well as Sun Solaris and HP UNIX users. For example, IBM’s Power enterprise 64-core server delivers twice the performance at a comparable price as a similarly configured HP Superdome Itanium® system. (1)
Very cool, I’ll need to read up about these new Power Systems, they do sound very cool. More choice in platforms and innovation has to be a good thing for the market and the end user in terms of choice and options for their business.
April 10, 2008 at 10:45 pm · Filed under blades, rackmounts
http://www.hp.com
GENEVA, Switzerland, Apr. 9, 2008 – HP today introduced four new operating environments (OEs) with the latest release of HP-UX 11i v3, Update 2. Designed to simplify UNIX deployments on HP Integrity servers, the new OEs enable customers to reduce costs and accelerate time to solution for immediate out-of-the-box productivity.
The OEs are HP-tested and pre-integrated to ease software management from purchase to installation to updates. They deliver fully integrated UNIX virtualization and high availability features, as well as a full range of HP software, with the HP-UX 11i operating system, in a single, easy-to-manage package.
The four OE offerings, Base, Virtual Server, High Availability and Data Center, allow any size business to rapidly and efficiently integrate a complete UNIX solution into its Integrity server environment. As a result, customers avoid potential errors and achieve faster time to solution by simplifying the manual effort typically associated with installing a fully-functional UNIX environment. For example, with the Data Center OE, customers can install more than 150 different software components in just nine steps—compared to the 35 steps required for IBM AIX and 87 for Sun Solaris.
Very cool, anything we can do to make deployment and and management in the UNIX arena has to be a good thing, I’ll need to read up more.
Next entries »
