Curtis Combs Jr. Email & Phone Number

Falls Church, Virginia, Us

Marietta, Ga, Us

San Jose, California, Us

Tampa, Fl, Us

I am currently the singular Systems Administrator of a small HPC cluster which utilizes PBS/Moab as the scheduler, CentOS 6 and 7 as compute nodes and provides compute capability for a mixed team of Bioinformaticians and Biostatisticians.

Presently, my renewed position with the University of Georgia includes the back up support of the existing cluster and secondary administrator, primary architect of the upcoming HPC cluster and duties as assigned by the lead administrator.

Davis, California, Us

In my short time at UCD I attempted to assist the current staff with operations for twelve HPC clusters for a variety of disciplines, all were Ubuntu-based clusters from 12.04 to 16.04 running the SLURM batch scheduler. The clusters were deployed using the Cobbler platform management system and maintained by Puppet. Storage was over NFS with MDADM RAID. Networking was 10 or 40Gbps Infiniband. When possible, I assisted with typical systems administrator duties. Mostly, I created user accounts and compiled software from source code. Please contact me for further details.

Statesboro, Georgia, Us

At Georgia Southern University, I am primarily responsible for the technical construction and administration of the Talon HPC cluster and I adopted the management of 3 additional clusters. Talon is created over reclaimed, grant-purchased and donated hardware with the intention of using a software stack that is completely open source. 53 nodes were managed using the SLURM scheduler, Lustre was used for the /scratch filesystem and NFS for /home, /work and /data mounts. Some of our nodes are virtual, some are physical. Authentication is based on LDAP. Multiple disciplines were permitted to run here. The hardware was diverse with memory ranging from 8GB to 256GB over all Xeon cores from Nehalem to Haswell.The second HPC cluster that I managed was a smaller Platform LSF based cluster, deployed with xCAT and running CentOS 6. The cluster has multiple 1Gb networking links and a 56Gb Infiniband network for MPI and inter-nodal traffic. 13 nodes with hardware configurations ranging from 60-core Ivy Bridge to AMD Opteron 6xxx.The third and fourth clusters which I adopted were loose collections of nodes with NFS mounts which I managed using simple tools like pdsh and collections of BASH scripts. One of these clusters, primarily used for comp. chemistry utilized CentOS with NFS over E5-24xx and E5-26xx Sandy Bridge, all 32-core/2-socket/2-thread.

Ca, Us

In this role, I worked with a Computer Science forensics postgraduate research laboratory, supporting their needs for system administration using VMWare vSphere 5.5, Fedora/CentOS-based Linux and Windows 2008/2012 servers. My day-to-day duties included making and managing student and faculty accounts, working with campus IT departments and vendors, designing virtualization solutions and assisting with High Performance Computing job submissions.

Athens, Ga, Us

At this position I was responsible for building our primary virtual infrastructure on VMWare 5.1 vSphere Enterprise. I installed, configured and was primarily responsible for managing the ESXi hosts in three HA zones, managed by vCenter. Storage was implemented over iSCSI from Dell Equalogics. The clusters supported around 20 virtual machines, most of which were mission-critical, some were ancillary. Our login nodes and platform management servers were there. We used the Rocks cluster management software for cluster deployment, PXE, TFTP and Kickstart. Our entire cluster was first RHEL 4, then RHEL 5 and we worked with some CentOS 6 nodes.I was also primarily responsible for storage management. Our HSC had fast, HPC storage on Netapp first, then Panasas, which I managed. Then, secondarily, we had 5 Solaris 10 SunFire units with ZFS. Then a third layer running OpenIndiana also with ZFS.I shared duties managing and maintaining the scheduler, which was Univa Grid Engine, a Sun Grid Engine variant. I shared this duty with a primary systems administrator but it was a very busy position. Personally, I configured scheduler weights to optimize cluster conditions for job dispatch. We regularly had BLAST, velvet, BLAT and refseq jobs in the thousands both running and queued. The cluster was a hybrid, mostly comprised of 1GB server-class Dell hardware but also many racks of QDR InfiniBand connected nodes, "large memory (512GB,1TB)" nodes, "many core (greater than 24-core)" nodes, and nodes with Keplar-based GPGPU CUDA cards.In this position, I primarily used BASH and Python for scripting and automating cluster load sensors, high-level job interpretations for our managers and automated installation routines. We used Ganglia and OpenNMS for monitoring the cluster and troubleshooting with SNMP. Our switch throughput was in Nagios, through our NOC and I was the network liaison for our networking requests.

Athens, Ga, Us

Note: This position is partially the same as the above position with the Georgia Advanced Computing Resource Center. This was a change of name when the institution moved under new management with the Institute of Bioinformatics at the University of Georgia from the Office of the Vice President.In my capacity at the Research Computing Center and Office of the Vice President, I worked 50/50 time. Mostly for the RCC I supported the Netapp storage systems, while performing System Administration duties for the Office of the Vice President. I managed the OVPR storage systems as well, since they were Netapp units, too.At the Office of the Vice President, I managed a VMWare installation starting on VMWare 3 and migrating to VMWare 4.1. The cluster was completely comprised of three large Dell R900 servers. These servers supported over 80 virtual machines, which OVPR applications were expressed and there were many roles which I both built and supported. As the primary administer, I worked closely with application developers, web designers and vendors. Our developers were primarily Tomcat developers that used MySQL databases. I worked closely with the development team to implement their server needs to support their development-to-production cycle and installed and configured SVN servers, development Tomcat 5 and 6 servers and MySQL database servers (we had three).I also worked with our web designer in the same capacity. The Office of the Vice President for Research hosted around 40 virtual hosts under the uga.edu SOA. I was responsible for assisting her by installing and configuring many instances of Wordpress and its associated plugins and databases. She also used other CMSes such as Expression Engine and MediaWiki. Each site required configuration of the Apache webserver to host the sites including development domains.There were other, smaller, applications which used clusters of Windows Server 2008 and SQL Server to support some departments under the OVPR.

Athens, Ga, Us

At the Anthropology department, I was responsible for mostly desktop support functions. I automated the installation of some desktop software. One major accomplishment was joining the entire department the campus Active Directory domain. There were around 100 computers total and I managed the software installation for the department as well. While I was there, I implemented network shares off of Samba and I managed two web servers for the Archaeology department. These web servers hosted the department's website for the Georgia Site Files and were based on Apache with MySQL databases.

Gainesville, Georgia, Us

Our responsibility was to provide support for 25 Public Health and Environmental Health departments in the 10 counties surrounding Athens-Clarke. I was responsible for desktop and server support as well as installing, planning punch-down closets and the few servers at each location (an OpenBSD-based VPN and a webserver). I personally managed and installed some T1 lines at the larger locations as well as constructed a VPN tunnel solution for every location to access the Headquarters' resources securely.

Valdosta, Ga, Us

My position as Multimedia Specialist at Valdosta State University was to assist professors in creating presentations for their classes using non-linear video editing tools, sometimes Powerpoint and Word. I used Adobe Photoshop daily. I used Adobe Premier daily. The Media Center at VSU was also responsible for printer repair for the entire campus and we often seen at least one printer per day that needed maintenance or repair. This was not bulk module repair, it was component repair using breakdown diagrams from the service manuals.

Bachelor’S Degree, Biology/Biological Sciences, General

College Prep, College/University Preparatory And Advanced High School/Secondary Diploma Program