Linux Systems Engineering
Deep Linux and virtualisation engineering, including the incidents that vendor support escalates rather than solves.
Linux is where most of the infrastructure we work on actually runs, and it is the layer where a shallow supplier becomes expensive. GİTA Teknoloji works across enterprise Linux estates — Red Hat, Rocky and AlmaLinux, Debian and Ubuntu, SUSE — from initial architecture through to the awkward production incidents that do not resolve themselves.
The work spans three areas: building and standardising Linux estates, operating and hardening them, and diagnosing failures in the layer beneath them — hypervisor, storage fabric and hardware. That last area is the one customers most often lack, because it requires reading counters at the HBA and multipath level rather than in an application log.
Our technical background is in this layer: system virtualisation, distributed storage, network virtualisation and clustering, including three granted patents and R&D work funded through TÜBİTAK.
Systems engineering
Architecture and build standards
Reference builds, partitioning and filesystem layout, package and repository strategy, so servers are reproducible rather than individually hand-made.
Hardening
Baseline hardening against recognised benchmarks, service minimisation, SELinux or AppArmor policy, SSH and access control, audit logging.
Performance analysis
Diagnosis at the level the problem actually lives — scheduler, memory pressure, I/O queueing, network stack — rather than adding resources and hoping.
Patching and lifecycle
Patch strategy, staged rollout, kernel lifecycle and end-of-life migration planned before it becomes urgent.
Automation
Ansible for configuration and change, so an estate is described in code and a rebuild is a procedure rather than an archaeology exercise.
High availability and clustering
Cluster design, quorum, fencing and failover behaviour tested under real failure rather than assumed from the documentation.
Virtualisation and platform migration
Licensing changes have made hypervisor migration a live question for a lot of estates.
Proxmox VE
Cluster design, storage integration, backup strategy and production operation — including estates moving off commercial hypervisors.
KVM and QEMU
Native KVM deployments where a full management platform is unnecessary, including tuning for I/O-heavy and latency-sensitive workloads.
XenServer and Citrix Hypervisor
Long-standing experience with the Xen platform, including participation in the Citrix ISV programme.
Migration from VMware
Assessment, target platform selection, staged workload migration and rollback planning, rather than a single high-risk cutover.
Storage integration
Fibre Channel and SAN, iSCSI, NFS, multipath and ALUA configuration, and the failure modes each of them has.
Capacity and consolidation
Right-sizing based on measured behaviour, so consolidation does not simply move the bottleneck.
Incident diagnosis and root cause
The problems that survive first-line support usually live below the operating system.
Storage path failures
Diagnosis of multipath and ALUA path loss, transport errors and the physical-layer faults behind them — CRC error counters, optics, cabling.
Fibre Channel and HBA analysis
Reading error counters at the HBA and fabric level to distinguish a failing transceiver from a controller or firmware issue.
Kernel-level investigation
Crash and hang analysis, dmesg and trace interpretation, driver and firmware interaction.
Reproducible diagnostics
Diagnostic tooling written for the estate, so the same check can be run again next time instead of reconstructed from memory.
Written root cause
A documented cause and remediation, because an incident that is fixed but not explained tends to recur.
Second opinion
Independent analysis where a vendor support case has stalled or the proposed fix does not match the evidence.
Ongoing operation
Managed Linux estates
Day-to-day operation under a defined support window, with monitoring, patching and change handled as one service.
Monitoring that reflects the estate
Checks built around how these systems actually fail, rather than a default template applied uniformly.
Backup and recovery
Backup design, and recovery tested to the point where the recovery time is a measured number.
Access and identity
Centralised authentication, privilege management and auditable administrative access.
Documentation
Runbooks and architecture documentation maintained as part of the service, not written once at handover.
Knowledge transfer
Working alongside your own engineers where the goal is capability in-house rather than permanent outsourcing.
Frequently asked questions
Which distributions do you work with?
Red Hat Enterprise Linux and its rebuilds — Rocky and AlmaLinux — along with Debian, Ubuntu and SUSE. The engineering below the packaging layer is largely common, and we work with what the estate already runs rather than pushing a preferred distribution.
Can you take over an estate we already run?
Yes. That usually begins with an assessment: what exists, how it is configured, where it diverges from itself, and what is undocumented. That assessment is worth doing on its own merits even if no ongoing service follows.
Can you help us move off VMware?
Yes. Proxmox VE is the most common destination and we run production estates on it, but target selection should follow an assessment of the workloads rather than a general preference. Migration is staged, with rollback planned.
Do you provide 24/7 support?
We offer a business-hours support window as standard, with extended coverage where the estate justifies it. Monitoring runs continuously regardless of the support window.
Will you work alongside our own team?
Frequently, yes. A common arrangement is that we handle architecture and escalation while the in-house team runs day-to-day operations, with documentation and knowledge transfer built into the engagement.
Do you do one-off incident work?
Yes. Independent diagnosis of a specific production problem — particularly where a vendor case has stalled — is a normal way for an engagement to start.
Related services
Have a problem that has outlived first-line support?
Independent diagnosis of a specific production issue is a good place to start.
Get in touch