In 2016 Ellexus launched its ground-breaking new tool for load balancing shared storage. Developed in collaboration with the IT department at ARM.
Always-on monitoring in production
Mistral is the leading I/O profiling tool in the new family of always-on monitors of Linux system. By allowing you to control the amount of information gathered, you can ensure that you know about the important events without being overwhelmed with too much data. It is storage agnostic and monitors I/O bandwidth, latency and meta data as well as system stats such as CPU and memory utilisation.
Mistral can be used with Elasticsearch and Grafana, or other frameworks such as Splunk, to give a system overview with per user, per host or per project overviews updated in real time.
Once a problem has been identified, the specific user, job, host, program and file can be extracted from the Mistral log data and the problem quickly solved.
Identify rogue jobs in production and test
Mistral is lightweight enough to run in production or on large test suites to find bad I/O before it becomes a problem.
Protect the storage by throttling job I/O
Mistral can actively protect shared storage by throttling the I/O of individual jobs to maintain a good quality of service.
Eliminate bad I/O patterns
Mistral can detect bad meta data accesses, random I/O and small reads and writes as well as simpler problems such as writing too much data. It gives users the data they need to fix the problem.
Category
Red Hat Certifications
This product has been certified to run on the following Red Hat products and technologies:
| Target Product | Level |
|---|---|
| Red Hat Enterprise Linux 7.x | Self-Certified |
| Red Hat Enterprise Linux 6.x | Self-Certified |
