On 3 May 2000, at 21:21, Robert L. Millner wrote:
mistake to have been. Was my mistake not buying an EMC or a Sun in the first place or was my mistake not noticing the autosupport message (for example)?
We have some big EMC systems - they aren't perfect either! I've had: - databases crash due to scsi bus errors that emc couldn't explain (it was an emc problem, the same exact scsi error occured on several systems at the exact same time) - 30 min outages for microcode updates that turned into 4 hours - lost a 4 port scsi board in the emc box taking systems down
Our emc sales rep insists that they've never lost data. Maybe . . . but I don't see where it's any less vulnerable to double- disk failures than any other raid system.
Rick
On May 04, Richard L. Rhodes wrote:
We have some big EMC systems - they aren't perfect either! I've had:
- databases crash due to scsi bus errors that emc couldn't explain (it was an emc problem, the same exact scsi error occured on several systems at the exact same time)
- 30 min outages for microcode updates that turned into 4 hours
- lost a 4 port scsi board in the emc box taking systems down
Our emc sales rep insists that they've never lost data. Maybe . . . but I don't see where it's any less vulnerable to double- disk failures than any other raid system.
For a short while, when we were very young, NetApp could honestly make the claim, "We have never lost any customer data". At our size now, that's simply impossible, and I know for a fact that it is also not the case for EMC.
I absolutely understand that some of our customers have lost data, and that others have been unable to access it for a period that was unacceptable to them.
I'm sorry.
If it's any solace, I'd like to let you know that we have learned a lot from your pain, and we are doing our best to reduce the odds that you -- or other customers -- will run into these same problems.
I also have a request. Please turn on autosupport. That's extremely important to us for two reasons:
(1) When we do encounter one of these really nasty problems, one of the first things we do is check all the autosupports for the system to help figure out what exactly happened, and to see whether we can anything that would have let us catch the problem early. We save autosupports forever, and in some cases there have been unrecognized "problem signatures" going back for years!
(2) When we find a new signature, we add it to our autosupport scanning, as Karl described, and in many cases we will automatically open a case and contact the customer in order to resolve the problem before it turns into a crisis. Running autosupport is a way that you can help us to help you.
Again, I apologize that our systems aren't perfect, and I assure you that data safety is our top priority.
Dave Hitz VP of Engineering and Co-Founder
-
Dave Hitz -
Richard L. Rhodes