Welcome to OmniTI Labs
OmniTI Labs is our repository for some of the projects we generate while working on stuff for clients, for our products, or just on things that we think are cool. The website is a way for us to give this code back to the community.
We at OmniTI have grown our business around understanding our customers' businesses. We like to think that, after a while, we know them as well as they do. This allows us to choose technology that is most appropriate for their current challenges and anticipate their future challenges. It just so turns out that over 95% of the technology we deploy is open source. Open source technologies allow us to make software do whatever our clients envision... it is a characteristic of software that removes all boundaries.
Open source has been a vital component in our success. Writing software to make development and operations work easier is not our business model; providing the best development and operations work in the world is. It just makes sense to make most of our tools open to the world.
Like all good open source efforts, we encourage outside contributions. Many of the projects that you'll find on this site have contributors outside of OmniTI and a growing number are maintained by the community. After all, that's the point... community.
Reconnoiter is making waves
Our efforts to revolutionize the trending of systems, applications and network metrics has been noticed by others. Reconnoiter has made great strides in both fundamental design and usability.
Zetaback wins award!
Zetaback is among good company as it takes third prize in the Open Solaris Community Innovation Awards.
News and Announcements
OmniTI Labs announces Zetaback release 1.0.6. This release includes a number of new improvements/bugfixes including: improved violator report, more informative error messages and a setting to limit the number of concurrent zetaback processes running.
OmniTI Labs announces Zetaback release 1.0.5. This release includes a number of new features including: the ability to back up directly to a dataset instead of to files, flexible retention policies, the ability to mark a host as offline so backups won't be taken while still allowing the listing and archiving of backups, and support for backup 'classes' allowing options to be set on a per-filesystem basis.
OmniTI Labs announces Zetaback release 1.0.4. This is primarily a bugfix release. Since version 1.0.1, ZFS volumes (zvols) have been skipped by zetaback_agent. Now zvols will be discovered like any other dataset. Users running ZFS root pools may wish to set explicit excludes on their swap and dump zvols to avoid having them backed up after upgrading to 1.0.4.
OmniTI Labs announces Zetaback release 1.0.3. This is primarily a bugfix release. Fixed bug where a failed incremental backup would cause all subsequent backups for that host to be skipped. Fixed a bug with the version string reported by the -v option. Clarified docs on the 'violators' view.
OmniTI Labs announces Zetaback release 1.0.2. New in this release are a config option, ssh_config, for specifying an alternate ssh config and support for excludes based on ZFS user properties. For more information, see the man pages for zetaback(1) and zetaback_agent(1), respectively.
OmniTI Labs announces Zetaback release 1.0.1. Highlights of this release are support for compressing backups, and a bugfix for Nevada releases >= 99 to accomodate a change in the default output of 'zfs list'. This fix is fully backward-compatible with previous versions of Solaris 10 and 11.
The latest entry to the OmniTI Labs stable is nothing new. For the past eight years, we've hosted a patch against portable OpenSSH that allows it to authenticate directly against RSA SecurID ACE server. OmniTI Labs is the perfect home for this effort. This marks the first Labs project that is maintained outside of OmniTI; since 2003, Jim Matthews has ensured that the SecurID patch works against the latest OpenSSH release.
Monitoring and trending isn't rocket science -- it's computer science. The most important aspect to understanding data is that you need the data before you have a problem -- you can't have an anamoly without a baseline. One of the challenges with data collection and graphing systems today is that it takes a while to add data points and new metrics into the system. Reconnoiter makes this easy. It strongly decouples data collection in a distributed manner, data aggregation and reporting so that each piece can be individually tackled with best-of-breed design.
What's my machine doing? It's over there and I can't see how many whizzies are whizumping because there isn't an snmp MIB for that. I'm too lazy to extend SNMP and I don't want to compile anything. Resmon's first goal: to be simple, the second goal: useful. It accomplishes both exceptionally well. Exposing obtuse metrics on a host has never been easier. It's over HTTP, uses XML. Getting the metrics remotely is trivial both in retrieval and digestion.
OmniTI Labs and the Zetaback team presents Zetaback version 1.0.0. After one year of heavy usage, Zetaback emerges from Alpha. To quote a Zetaback team member: "This tool has saved my ass so many times I've stopped counting."