I have about 20 HP DL 380 (some 360) servers, from Gen7 to Gen9, using the HPSA driver with various smartarray controllers. For a long time I've been running mainline 3.14 kernels, without any issues. Some time ago I updated to mainline 4.4.x, up to the most recent 4.4.30. Now I noticed, especially on one server, but in the logs on 6 of them, the following kind of message: 2016-11-06T22:09:50.227592+01:00 HOST kernel: [68853.338610] hpsa 0000:03:00.0: scsi 0:1:0:0: resetting logical Direct-Access HP LOGICAL VOLUME RAID-5 SSDSmartPathCap- En- Exp=1 2016-11-06T22:10:18.713759+01:00 HOST kernel: [68881.832436] hpsa 0000:03:00.0: scsi 0:1:0:0: reset logical completed successfully Direct-Access HP LOGICAL VOLUME RAID-5 SSDSmartPathCap- En- Exp=1 I see such messages, _usually_ only with 1 second between resetting/reset, on machines with the following controller+controller firmware variants: 1 P410i 5.14 1 P420i 5.42 2 P440ar 3.02 1 P440ar 3.56 1 P440ar 4.02 The one machine for which I've shown the concrete message, is a P440ar with firmware 3.02. There, contrary to the other machines, it sometimes takes up to 20 seconds for that resetting operation, and meanwhile, all I/O stalls. I also tested with 4.8.x kernels, and saw the same symptoms there. I'm somewhat sure that I did not see these with 3.14 kernels. This morning I rebooted the most problematic box to 3.14.79, so far it was silent. I'll report if that changes. Apart from these log lines, there is nothing strange to be found - no ILO or IML notifications visible, no other kernel messages, no drive failures, SMART alerts, or performance regressions...
Some more info on my problematic machine / further diagnosing is in https://bugzilla.kernel.org/show_bug.cgi?id=187231 Summary: at least with the P440ar controllers, such 10-30 second "logical reset" episodes eventually reveal an underlying faulty drive, and go away when that is drive is replaced. But there is no up-front information in the "logical reset" that would permit pinpointing the drive on the first round.