BOSTON- The Red Hat Enterprise Linux paradox is that it gets deployed because companies see it as “the” corporate Linux distribution, with credible support. But in the past, many RHEL users said support that looked fine on paper often fell short once a call was made.
That’s changing, according to attendees at the Red Hat Summit 2011 conference.
Red Hat support has improved significantly over the past few years, said Munjeet Singh, senior associate of dynamic infrastructure at Booz Allen Hamilton, a consultant supporting the U.S. Army’s data center consolidation effort.
One reason is that the current RHEL 6 release includes a customer portal and the ABRT (automated bug report and tracking) tool that automatically collects data about the nature of a problem so the support team gets a more detailed picture of what went wrong. That can eliminate a lot of the frustrating back-and-forth between customer and vendor support staff and also closes the loop with product developers looking to improve the code.
Red Hat portal, ABRT fills support data gap
Insufficient information about the tech snafu can cause huge issues, said Jerry Levy, a support engineer with a large storage company. Levy often has to work with other vendors’ support staff to help customers solve problems in multi-vendor data centers.
To minimize hassle, users should work with their support engineers, Levy said. “We need to know the date and time of the problem so we can locate it in the logs. When the customer provides that, the process goes a lot faster. What’s really death is when you call [support] because someone else told you to and you have no idea what the problem was or when it occurred,” Levy said.
Any tool that automatically aggregates problem data and makes it available electronically to vendor support is a very good idea, he said.
Both the portal and ABRT will be enhanced the upcoming RHEL 6.1 release. For example, in 6.1, the email notification will be on by default, said Andrew Hecox, engineering manager at Red Hat.
RHEL 6.1 will also add “faceted search,” so searches can be filtered or narrowed to focus just on the kernel or just middleware, said Hecox.
The current ABRT iteration (called “abort” by Red Hat support staff) provides helpful data on C or C++ crashes, kernel “oopses” and unhandled Python exceptions, Hecox told attendees at a panel session this week.
RHEL 6.1 version will add enhancements and fixes and subsequent releases will provide customers near real-time suggestions for possible fixes even before they fill out a support ticket, said Hecox,
Many of the improvements will come because Red Hat is logging huge amounts of data from customer problems, stripping out the customer name, and distilling them into KnowledgeBase articles that can be used to solve future issues, said, Marco Bill-Peter, Red Hat vice president of global support services. Post RHEL 6.1, the customer portal will be able to help users even as their problem data flows into the system. “An algorithm will match symptoms and suggest solutions even before a ticket is submitted,” he said.
Some Red Hat support gotchas persist
Still, there are nagging issues. A Midwestern Red Hat partner who supports his customers said the quality of “level one” support at Red Hat remains inadequate in his view. “Today, the level one guys you talk to are all but clueless and unless you can get to the level two people you’re out of luck,” he said.
Strictly speaking, Red Hat, like most tech companies, eliminated the formal level one through four support structure but in practice most customers and vendor staff still use that terminology.
And for all the streamlining the company has done, there are still glitches.
This partner recently helped a customer set up a big, four terabyte Dell server running Oracle on RHEL.
“I called Red Hat, requesting a white paper that my account will not let me access. I was told they had technical support problems with the partner website, and would have to wait.”
Since the customer was ordering the hardware immediately, the partner called a dozen Red Hat people trying to get the documentation. He failed.
“Instead, we just took a SWAG, (a silly wild ass guess)…and I was able to get hardware configured and installed but it was tenuous, we would have had more confidence when we placed the order with Dell if we had had the white paper in hand,” he said.
Let us know what you think about the story; email Barbara Darrow, Senior News Director at firstname.lastname@example.org.