Twice now on a cold restart of my Test Bed server (spec's below), I have had the cache drive come up missing when I restarted the server.   The Drive is detected  (WD 320GB  S/N  6353 (sdd)  ) as an unassigned drive.  When I stopped the array, I can assign that drive as the cache drive and on restarting the array everything works fine.  Below is a section of the syslog which shows that the drive is detected but is not BUT assigned to be the cache drive. This section of the syslog shows that the drive was detected and later,   the    import 30 cache device:  no device   statement.

 

Jul 26 12:42:02 Rose emhttpd: shcmd (114): udevadm settle
Jul 26 12:42:02 Rose emhttpd: Hitachi_HDS721010CLA332_JP2911HD0Y80KC (sdg) 976762552
Jul 26 12:42:02 Rose emhttpd: WDC_WD30EFRX-68EUZN0_WD-WMC4N0K8P3JK (sdh) 2930266532
Jul 26 12:42:02 Rose emhttpd: WDC_WD3200AAJS-56M0A0_WD-WCAV2L226353 (sdd) 312571192
Jul 26 12:42:02 Rose emhttpd: Hitachi_HDS721010CLA332_JP6930HD15DEMF (sde) 976762552
Jul 26 12:42:02 Rose emhttpd: err: ckmbr: No medium found (123): open: /dev/sdb
Jul 26 12:42:02 Rose emhttpd: device /dev/sdb problem getting size
Jul 26 12:42:03 Rose emhttpd: Hitachi_HDS721010DLE630_MSK5235H2JJTTL (sdf) 976762552
Jul 26 12:42:03 Rose emhttpd: Hitachi_HDS721010CLA332_JP6930HD13JA2F (sdc) 976762552
Jul 26 12:42:03 Rose emhttpd: Generic-_USB3.0_CRW_-0_000000015140-0:0 (sda) 7639008
Jul 26 12:42:03 Rose kernel: mdcmd (1): import 0 sdh 2930266532 0 WDC_WD30EFRX-68EUZN0_WD-WMC4N0K8P3JK
Jul 26 12:42:03 Rose kernel: md: import disk0: (sdh) WDC_WD30EFRX-68EUZN0_WD-WMC4N0K8P3JK size: 2930266532 
Jul 26 12:42:03 Rose kernel: mdcmd (2): import 1 sdg 976762552 0 Hitachi_HDS721010CLA332_JP2911HD0Y80KC
Jul 26 12:42:03 Rose kernel: md: import disk1: (sdg) Hitachi_HDS721010CLA332_JP2911HD0Y80KC size: 976762552 
Jul 26 12:42:03 Rose kernel: mdcmd (3): import 2 sde 976762552 0 Hitachi_HDS721010CLA332_JP6930HD15DEMF
Jul 26 12:42:03 Rose kernel: md: import disk2: (sde) Hitachi_HDS721010CLA332_JP6930HD15DEMF size: 976762552 
Jul 26 12:42:03 Rose emhttpd: import 30 cache device: no device
Jul 26 12:42:03 Rose kernel: mdcmd (4): import 3 sdf 976762552 0 Hitachi_HDS721010DLE630_MSK5235H2JJTTL
Jul 26 12:42:03 Rose kernel: md: import disk3: (sdf) Hitachi_HDS721010DLE630_MSK5235H2JJTTL size: 976762552 
Jul 26 12:42:03 Rose kernel: mdcmd (5): import 4 sdc 976762552 0 Hitachi_HDS721010CLA332_JP6930HD13JA2F
Jul 26 12:42:03 Rose kernel: md: import disk4: (sdc) Hitachi_HDS721010CLA332_JP6930HD13JA2F size: 976762552 

The virtually identical setup for the Media server works just fine the few times that it has been restarted.  I have attached the complete diagnostic file from after the server was rebooted that I generated, and the last diagnostics file that the Fix Common Problems saved to the logs folder on my Flash Drive. 

 

Anyone have any thoughts as to what is going on?

rose-diagnostics-20170726-1247.zip

rose-diagnostics-20170726-1224.zip

Edited July 28, 20178 yr by Frank1940

23 minutes ago, Frank1940 said:

Twice now on a cold

 

By twice do you mean you only rebooted twice and it happened both times or you rebooted more times and it only happened twice?

I have only done it twice since installing the cache drive and it has happened both times.  I wish to clarify that both times was a startup from a power-off state not a reboot (auto stop and restart of the system without a power shutdown).

 

First time, I assume that it was some sort of gremlin but when it happened the second time, I decided that I needed some help in trying to sort it out.

Looks like a bug and probably related to the import cache line coming up before all the disks, it's usually at the end:

 

Jul 26 12:42:03 Rose kernel: md: import disk2: (sde) Hitachi_HDS721010CLA332_JP6930HD15DEMF size: 976762552 
Jul 26 12:42:03 Rose emhttpd: import 30 cache device: no device
Jul 26 12:42:03 Rose kernel: mdcmd (4): import 3 sdf 976762552 0 Hitachi_HDS721010DLE630_MSK5235H2JJTTL

IMO you should repost as a defect report.

I requested that it be transferred.  Thanks for your effort!

I forgot to mention but since you are a regular forum user I assumed you already tried both options on the FAQ that can fix this, if not try them before posting a defect:

 

 

If johnnie.black's suggestion doesn't help

 

1 minute ago, Frank1940 said:

I requested that it be transferred.  Thanks for your effort!

 

I will move this thread to Defect Reports if you will edit the first post to conform to this:

 

@johnnie.black, I tried both suggestions and used IE and FireFoxESR (52.2.1) and made the change assignment in the safe mode. Either suggestion solved the problem

 

@trurl, I have changed the subject line.

Edited July 27, 20178 yr by Frank1940
word change for clarity

This problem has been solved for me with the release of version 6.4.0-rc7a.  It did take a couple of reboots to get everything back in sync.  (That is the best way to describe what I saw happening...)