Tube-Mac Industries, a multi-location Ontario manufacturer, needed infrastructure that stayed up, email that belonged in this decade, monitoring that noticed problems before people did, and a better way to manage jobs against its Sage 300 accounting system.
We deployed Nimble enterprise storage under a VMware environment engineered for high availability across sites, took on proactive monitoring and support, migrated to Microsoft 365, and built JobSight, a custom job management application integrated directly with Sage 300.
Manufacturing punishes downtime differently from office work. When an accounting system is unavailable, invoices go out late. When the systems behind production are unavailable, people stand still and material sits where it was put. The cost is not proportional to the length of the outage, because restarting is not free.
Tube-Mac operates from multiple locations, which changes the problem in a way that is easy to underestimate. A single site can be made reasonably resilient by adding redundancy inside one room. Multiple sites that depend on each other's systems raise a second question: what happens to the other locations when one of them has a bad day? Resilience has to be designed across the estate, not purchased separately for each part of it.
Alongside the infrastructure work sat an operational irritation tolerated for a long time. Job information lived in one place, financial information lived in Sage 300, and the connection between them was people. The same details were entered more than once, which costs time on the way in and costs confidence on the way out, because when two systems disagree somebody has to decide which one is lying.
Storage first, then virtualization, then the operational layer, then the software the business had been asking for.
Shared enterprise storage is the precondition for meaningful high availability in a virtualized environment. Until the disks live somewhere independent of any one server, a workload cannot move between hosts without moving its data too, and high availability quietly reduces to a promise about how quickly somebody can restore from backup.
Virtualization was designed around the failure cases rather than around the capacity numbers. Workloads are not tied to a specific host, hosts can be taken out of service deliberately for patching, and the multi-location design makes resilience a property of the whole estate rather than of whichever site happens to have the newest hardware.
High availability protects against components failing. It does nothing about the slower failures — a volume filling, a backup quietly failing, a host degrading — which are the ones that actually cause outages, because they are visible long before they become urgent. Continuous monitoring turns those into scheduled work instead of emergencies.
Mail and collaboration moved onto a current platform, which also removed a mail server from the list of on-premises things that must never break.
JobSight is a custom job management application that talks directly to Sage 300. It was built around how this manufacturer actually runs jobs and integrated at the accounting boundary, so job data and financial data are the same data rather than two versions of it. The double entry absorbing staff time was removed at the source, not automated around.
The last row is the one a generic infrastructure project would have left on the table. It was also the one staff noticed first.
High availability is a design property, not a product. It is entirely possible to buy expensive redundant hardware and still have an environment where one failure stops work, because redundancy inside a chassis does not help if the workload cannot run anywhere else. Three questions are worth putting to any proposal: which specific failures does this survive, what happens to the other sites, and how do you patch it without an outage. Vague answers indicate a vague design.
The integration point is what most manufacturers under-invest in. Every plant has a place where the shop floor's understanding of a job meets accounting's understanding of the same job, and in most businesses that place is a person retyping. That work is invisible in a budget because nobody bills for it, but it is expensive, and it is why two systems can both be current and still disagree. Off-the-shelf job management assumes a way of running jobs; when it fits, buy it. When it does not, changing your process to suit the software is usually the more expensive of the two options available.
Related work: manufacturing IT services and hyperconverged computing.
Ask us for a straight answer about which failures your current environment actually survives, or about integrating a job system with Sage 300. We target a 15-minute response on critical issues.
%20(1).avif){kind=logo}
%201.avif)
