He should have called Zerowait for his NetApp support!

NetApp’s midrange storage quality up for debate
By Beth Pariseau, News Writer
22 May 2006 | SearchStorage.com

It may have been a blow to Network Appliance Inc. (NetApp) that its traditional sweet spot, enterprise NAS, went to Hitachi Data Systems Inc. in Storage magazine’s most recent Quality Awards survey. (See HDS wins enterprise NAS, May 3.) But NetApp still made a strong showing in midrange NAS among survey respondents, a fact which baffled one midrange user who heard the results of the survey announced by Storage editors during last week’s Storage Decisions conference.

“I was amazed when I saw that rating,” said Nick Suranyi of Press Ganey Associates, which processes patient satisfaction surveys for hospitals and healthcare providers.

Ganey has 20 terabytes of NetApp midrange filers and needs to add about 20 more terabytes of storage this year as new regulations in many states are forcing healthcare providers to conduct the kind of surveys Suranyi ‘s company stores and processes. Suranyi himself said he is new to the company and inherited the FAS270s, R100s and FAS810s that compose most of Ganey’s environment.

“We’re seeing disk failures running rampant across every one of our filers,” Suranyi said. “We’ve had disks failing as often as two or three times a week — now it’s down to about once every three weeks, but for awhile there my desk was like shipping and receiving with replacement disks coming in.” Some of the crashes, he said, have resulted in data loss and system corruptions requiring more repair than simply replacing one drive.

Suranyi said he is confident he’s ruled out a version problem with his OnTap OS, or a problem with his environment in part by comparing notes with former colleagues and acquaintances in a test lab at Notre Dame University located in Indiana.

“I’m pretty sure, at this point, there’s something wrong with the way NetApp’s filers identify which disks are failed,” he said.
*****Perhaps he could use our ZHA Exception reporter?*****

And, he said, even if configuration problems were causing disks to crash, he’d never seen it happen so often or with such disruptive results.

“I’ve had Dell [Inc.] DA PowerVault 220s running for three years without seeing a failed disk,” he said.

Worse, he said, when the disk fails, if the filer does not detect a spare, it shuts the entire array down. And NetApp has repeatedly failed to meet its service level agreements (SLA) in response to the drive failures, often taking as much as three times the four-hour agreement time to deliver and install a new drive.

“When I complained to NetApp about it,” Suranyi said, “they told me the shutdown of the array was a feature to force me to protect the data better. And their only response about failing to meet the SLA was to do it again.”

He shook his head. “I couldn’t believe it. And how are they No. 1 in the midrange when they have all these problems? These are midrange filers we’re using.”

Suranyi said he plans to replace the NetApp filers completely this year with a SAN, probably from Xiotech Corp.

Meanwhile, however, Suranyi was overheard by James Jancewicz, storage administrator for a large health insurance provider based in New England, who asked that his company not be named.

“We love NetApp,” he chimed in. “It’s very transparent from the user’s perspective,” but he admitted that his company does see disk failures about once a month. “But six drives a quarter in half a petabyte isn’t all that bad, and we can fix it quickly,” Janceqicz said.

“To me, that many drive failures is unacceptable for an array,” said Suranyi.

“We also have yet to do a code upgrade on NetApp,” Jancewicz also conceded. “Whenever we buy a NetApp box, we just leave it there. We’ve had some of their boxes for three years without touching them.”

Jancewicz added, he’d received “very good support” from NetApp in general, a comment that surprised Suranyi.

“Look at what kind of company he works for, though,” said Bonnie Reif of LaPorte Hospital, herself an EMC user, pointing to the household name on Jancewicz’s conference name badge.

“That’s why it’s strange that they won for the midrange,” Suranyi said. “I had worked for some value-added resellers before that were also bigger and never experienced these problems.”

NetApp did not return calls for comment by press time.

This entry was posted in Uncategorized. Bookmark the permalink.