Port Multiplier Support?

[bbddpp]

My external enclosure connected via a port multiplier card seems to not be playing well with FreeNAS. Any suggestions on what I can try here? The entries in the log here are when I turn the enclosure's power on and off. I did that twice here. In previous versions, a power cycle at least allowed FreeNAS to then see the volumes in the enclosure. Now that I've upgraded to the latest, it doesn't seem to even be able to see the enclosure at all anymore.

Is an external eSATA enclosure connected to a port multiplier card not supported? Should I try USB?

EDIT: Running a /sbin/camcontrol rescan all allowed me to again see the drives in the enclosure and re-add them into the pool with an auto import, though I'm looking for these volumes and shares to survive a reboot.

Jun 26 17:01:33 freenas kernel: ugen3.2: <Generic> at usbus3 (disconnected)
Jun 26 17:01:33 freenas kernel: ukbd0: at uhub3, port 1, addr 2 (disconnected)
Jun 26 17:01:33 freenas kernel: uhid0: at uhub3, port 1, addr 2 (disconnected)
Jun 26 17:01:51 freenas kernel: (pmp0:ahcich1:0:15:0): lost device
Jun 26 17:01:51 freenas kernel: (pmp0:ahcich1:0:15:0): removing device entry
Jun 26 17:02:09 freenas kernel: pmp0 at ahcich1 bus 0 scbus1 target 15 lun 0
Jun 26 17:02:09 freenas kernel: pmp0: <Port Multiplier 38261095 1706> ATA-0 device
Jun 26 17:02:09 freenas kernel: pmp0: 300.000MB/s transfers (SATA 2.x, NONE, PIO 8192bytes)
Jun 26 17:02:09 freenas kernel: pmp0: 6 fan-out ports
Jun 26 17:02:29 freenas kernel: ahcich1: Poll timeout on slot 0 port 0
Jun 26 17:02:29 freenas kernel: ahcich1: is 00000000 cs 00000001 ss 00000000 rs 00000001 tfd 50 serr 00000000 cmd 0004c017
Jun 26 17:02:29 freenas kernel: (aprobe0:ahcich1:0:5:0): Unexpected signature 0xffff
Jun 26 17:05:44 freenas kernel: (pmp0:ahcich1:0:15:0): lost device
Jun 26 17:05:44 freenas kernel: (pmp0:ahcich1:0:15:0): removing device entry
Jun 26 17:06:18 freenas kernel: pmp0 at ahcich1 bus 0 scbus1 target 15 lun 0
Jun 26 17:06:18 freenas kernel: pmp0: <Port Multiplier 38261095 1706> ATA-0 device
Jun 26 17:06:18 freenas kernel: pmp0: 300.000MB/s transfers (SATA 2.x, NONE, PIO 8192bytes)
Jun 26 17:06:18 freenas kernel: pmp0: 6 fan-out ports
Jun 26 17:06:38 freenas kernel: ahcich1: Poll timeout on slot 0 port 0
Jun 26 17:06:38 freenas kernel: ahcich1: is 00000000 cs 00000001 ss 00000000 rs 00000001 tfd 50 serr 00000000 cmd 0004c017
Jun 26 17:06:38 freenas kernel: (aprobe0:ahcich1:0:5:0): Unexpected signature 0xffff

 

[monogrant]

Did you ever find a solution to this? I have the exact message using a 4 bay San Digital MS4UM+B

I gave up last night and switched to USB, but I only have USB 2.0 vs eSATA for four disks to share would be nice to get the port multiplier working.

 

[jgreco]

No. Port multipliers are dodgy and temperamental, and generally the SATA chipsets that aren't totally failtastic with PM's also have poor FreeBSD support, so it's just bad all around.

 

[Ericloewe]

I'll add that Intel explicitly states that its SATA controllers do not support port multipliers.

 

[mav@]

There are different SATA controllers. Some of them support port multiplers, other don't. Since in this case OS even tried to search for port multiplier, I can guess that the controller pretends to support it, but something is not going well. PMPs is an attempt to build something complicated for cheap market -- it is doomed for failure, since nobody wants to spend much money testing hardware for compatibility, updating firmwares, fixing bugs, etc. I can't say that FreeBSD's support for PMPs is guilty here more then hardware vendors. There are some examples when people like Backblaze managed to make PMPs work for them, but those are quite rare. Even much more expensive SAS expanders market has some problems when things get complicated enough, despite much bigger efforts to make it work.