Five years have passed and it is time to renew the storage. What are the most common errors?

The renewal of enterprise storage systems is a critical, complex and costly decision that should be approached with great care. It is common for mistakes to be made when planning and carrying out this type of project and in this article, we will discuss those that we encounter most frequently in our new clients.
in our new clients
.

1. Incorrect sizing

One of the most common mistakes when renewing enterprise storage is to underestimate or overestimate the capacity (in TB, throughput and response time) of storage needed. Incorrect sizing can result in underinvestment or overinvestment in enterprise storage. We all know that an NVMe solution is faster than an SSD solution, but do we need it? And for 100% of the data? At what price? Are the servers capable of reaching the bandwidth allowed by the cabin? (hint: in 95% of the cases no). But of course, in order to know this, it is advisable to carry out a previous sizing or audit, which has a very small cost in comparison with the acquisition of the cabins and what we can save. Hurries and offers for tomorrow are not good traveling companions in this type of projects.

Failure to forecast the cost for at least 5 years.

One of our customers purchased cabins from one of our favorite manufacturers years ago. It was extremely powerful and the price was very attractive. Two years later, they wanted to add records (they bought it with the bare minimum) and the list price of the records exceeded one million euros. Moreover, all the exercises showed that it was more economical to buy a new, lower-end cabin with more than enough performance and storage than to expand the existing one. Absurd? No, a very common commercial malpractice among companies without their own technical staff. Just as we get printers for free, similar things happen with many other HW components. Switches that are licensed per activated port, firewalls whose price varies according to the users that connect and thousands of other tricks to hide the costs in the medium and long term, as well as the dependencies generated and the obligation to buy additional servers and cabins every few years. Although all manufacturers do it, with the right advice it is often possible to turn it around and at the time of purchase to get at a very interesting price everything we need while the machine works.

3. Not knowing the life cycle of the machines (how long they will be maintained).

Another common mistake is not understanding the life cycle of the machines and not knowing how long maintenance services will be available. It is essential to take into account the useful life of the disk cabins and the availability of technical support and maintenance services in the medium and long term. Many times systems at the end of their marketing cycle are heavily discounted… but also during their launch. The latter allows us to use them for a much longer period of time, significantly reducing the annual cost. If we buy a booth that is going to be withdrawn from marketing soon, it may be that when we have put it into operation it will only have the minimum years of support left and after 5 years the only alternative is to support it ourselves, order the failed parts from international brokers or buy a new one. If we have purchased a system that has been on the market for a short time (but long enough for others to have tested it and experienced its first failures) it will be easy to extend the maintenance for 6, 7 or even 10 years, which is usually the top of its usual life cycle. Sales people from non-technical companies will tell you “you couldn’t tell” or that “it’s IBM / HP / Dell / Hitachi’s fault” but it’s not true. We have to know what we are selling and inform our customers in an honest way about the pros and cons of each new infrastructure we propose to acquire.

4. Not knowing the compatibility matrices with the operating systems.

Another common mistake is not considering the compatibility of enterprise storage with existing operating systems. It is essential to ensure that the storage is compatible with existing operating systems and that if we must upgrade, we do so prior to the implementation of the new storage. This is especially relevant when operating systems are out of support due to lack of preventive maintenance service. Maybe the new booth supports Red Hat or Windows, but not the production version because all new hardware requires drivers that would rarely exist before. Time travel doesn’t exist for operating systems either, and a 2008 Windows is likely to have compatibility issues with a 2023 cockpit.

5. Failure to anticipate the impact of migrations from existing production environments.

Finally, another common mistake is not anticipating the impact of migrations of existing systems on the storage we have renovated. It is important to carefully evaluate the existing systems to be migrated and understand how they will be integrated into the new storage solution, who will handle these professional services and what impact they will have on our organization or business.

6. Not having a trained technical team or L2-L3 support to call upon.

Many customers buy storage systems but do not have the time to be trained on them and do not have external technical support services to turn to, thinking that in case of any problem they can turn to the manufacturer because they already pay for support for these systems, when this is not always the case, especially if the system is working properly and the problem, although closely related, is something else.

7. Relying on companies that do not have technical teams and whose business is based on high margins on SW and HW sales.

Unlike our competitors, at Sixe Ingeniería we will always be on the side of our customers, because we live mainly from training, consulting and technical support. Services that we complement with the sale of HW that we normally help to maintain, so we are the main interested parties in ensuring that the solutions are ideal in the medium and long term. In addition, as a reference partner of IBM and Lenovo, we can offer you significant discounts, we know most types of critical environments very well and we are technology agnostic, although we have our preferences based on our experience. Contact us!

With IBM’s purchase of Red Hat, the entire portfolio of software solutions for cybersecurity, applications, databases, automation and hybrid cloud management has been ported to OpenShift under the brand name
IBM Cloud Paks
. This means that many of these applications have been redesigned and adapted to run on top of containers (although some, such as QRadar, have been doing so for years) and be controlled by Kubernetes, which is the container orchestrator on which OpenShift is based.

How are IBM Cloud Paks deployed?

IBM Cloud Paks are installed on a PaaS environment with OpenShift both in own data centers on IBM Power Systems, VMWare, KVM (RHEV / LinuxONE) and in public clouds of Microsoft (Azure), IBM, Amazon (AWS) and Google (GCP). Thanks to IBM Satellite, it can be deployed in a combination of on-premises and cloud resources, through a flexible hybrid architecture. Our
professional services department
can help you.

How are IBM Cloud Paks licensed? How much do they cost?

This is perhaps one of the least known and, in our opinion, most controversial parts. IBM has always sold perpetual licenses for all its software solutions. These licenses come with a basic technical support for HW incidents and a more advanced one for SW (SWMA) that can be renewed typically every 3 years. By moving to a cloud environment we are moving towards pay-per-use systems, which are very scalable and sometimes, to be honest, very expensive. complicated to estimate. For example, Data has these prices per “virtual core”. That is, from a few hundred dollars to a few hundred thousand… :)

This has obvious advantages for solutions where it makes sense to keep renewing support, as is the case with open and extremely complex solutions such as those based on micro-services and containers. Customers who are not comfortable with this model can continue to purchase appliances or licenses to deploy many of these solutions in their infrastructure for a one-time payment and optional support to be renewed every several years. In others, this is the only model as they are native solutions for Kubernetes and Cloud based environments.

Do I need to have OpenShift to install an IBM Cloud Pak?

Short answer: yes. However, if you don’t have it, you can deploy it without too many problems thanks to the installers included in the latest versions both in your own infrastructure and in an external one (IaaS) from your favorite cloud provider.

Are IBM Cloud Paks worth it?

As a good free verse that we are in the systems integrators sector, we think that some do, and others not so much (at least for now). It depends on their intended use, the dependency we have with other applications and the level of maturity in the adoption of containers and the use of Kubernetes in our organization. If we are starting out with Dockers, OpenShift and Cloud environments, perhaps it is better to stick to a good
digital transformation and modernization plan
rather than “putting the cart before the horse”.

Are there IBM Cloud Paks courses or training?

In order to take advantage of this technology you need to master both the infrastructure (OpenShift) for which there is official training offered by Red Hat, and an intensive hands-on workshop developed by ourselves. Once the infrastructure is under control you will need to train on the different IBM products and solutions you are interested in, as they are collections of Software grouped by category and licensed together. The Cloud Pak for Security, for example, is primarily IBM QRadar SOAR Platform while the Cloud Pak for Applications includes the entire Websphere suite.

That said, if you want,
we speak
.

We often overlook the need to perform preventive updates not only of the operating system (AIX, Linux, IBM i) but also of the FW of IBM Power servers. IBM publishes fixes for problems that other customers have experienced and it is usually not necessary for you to do so as well. In doing so, we keep our systems secure from all types of external and internal threats and vulnerabilities.

The problem we talked about in this short article is a bug that prevented LPARs from booting on Power9 servers if they had been running for more than 814 days. It sounds a bit like the printers of a few years ago that failed to print several hundred thousand pages, we never know if on purpose or by mistake. In the case of IBM it is a recognized firmware bug that is fixed with the update VH950_045_045 / FW950.00 available from November 23, 2020. So if you are an IBM customer where in the last two years, your aging Power9 systems have not been upgraded, you are likely to have this problem for the remainder of the year.

We give you a hint, the error is CA000040 which prevents the LPAR from booting and whose temporary solution could be to use the Power8 compatibility mode from the HMC or ASMI while you install the pending updates.

At Sixe Ingeniería we have been preventively maintaining our clients’ IBM Power systems for more than 15 years.
We can help you monitor, update and preventively maintain your entire infrastructure.
IBM and Lenovo server and storage systems. We also help you minimize licensing costs and offer the best prices on upgrades to new generations of IBM Power servers. Contact us for more information.