Please add function

[oler2]

Please add function Locate disk on LSI\ADAPTEC\HP HBA with WebGui

 

[Ericloewe]

Please note that feature requests have the most impact when filed in the bug tracker.

That said, this specific request is unlikely to gain much traction. For starters, Adaptec or HP HBAs are not suitable for TrueNAS. For LSI cards, you should typically be able to use the sesutil utility to do this from the CLI with minimum fuss:

1. Identify the device ID (e.g. da2)
2. Run (sudo) sesutil locate da2 on/off to turn on/off the LED for da2

Naturally this needs a compatible backplane, etc.

Also, please keep in mind that this is one of the features that iX prefers to reserve for its TrueNAS Enterprise systems, which have known hardware and for which the devs can ensure that such functionality works correctly and reliably. Supporting this could conceivably be easy on, say, Supermicro systems, but the nature of the community is that people tend to have... let's say quirky systems in most cases.

 

[sretalla]

Also, it's super-easy to locate any disk that has activity lights by using dd...

dd if=/dev/da2 of=/dev/null bs=1024k count=5000 This reads about 5GB of data from that disk (which should light its activity light for some seconds) and puts it nowhere. you could also use vibration or sound to identify the disk at that time if you have no lights.

 

[Ericloewe]

Also, it's super-easy to locate any disk that has activity lights by using dd...

dd if=/dev/da2 of=/dev/null bs=1024k count=5000 This reads about 5GB of data from that disk (which should light its activity light for some seconds) and puts it nowhere. you could also use vibration or sound to identify the disk at that time if you have no lights.

This, of course, is the ghetto option that is both inelegant and most compatible. Whether this is better or worse is left to the viewer's interpretation.

 

[Patrick M. Hausen]

Also, please keep in mind that this is one of the features that iX prefers to reserve for its TrueNAS Enterprise systems, which have known hardware and for which the devs can ensure that such functionality works correctly and reliably. Supporting this could conceivably be easy on, say, Supermicro systems, but the nature of the community is that people tend to have... let's say quirky systems in most cases.

I remember doing some tests - one time with sesutil, the other time with dd - and then rewiring my backplane to make them match.

 

[danb35]

sesutil did something strange to my system a couple of years ago, putting it into a very unstable condition that was only (ultimately) resolved by a complete power-off. I've always been a little cautious of it since then.

 

[Ericloewe]

sesutil did something strange to my system a couple of years ago, putting it into a very unstable condition that was only (ultimately) resolved by a complete power-off. I've always been a little cautious of it since then.

Huh, I've run it on my production NAS (Supermicro 1S something-or-other) a couple of times without any obvious issues. Any details you remember from your experience?

 

[danb35]

It's been a few years. Hardware was what's in my sig (probably different disk configuration, but same motherboard/HBA/chassis). What I remember is that I needed to replace a disk, so I ran sesutil locate da18 on (obviously substituting the right disk number) to locate the disk, and then removed the disk. But the light was still flashing, I wanted to turn it off, but da18 no longer existed, so I couldn't do sesutil locate da18 off (well, I could, but I got an error). So to turn it off, I instead ran sesutil locate all off. That did turn the light off, but (as I mentioned above) it left the system very unstable, in that it wouldn't stay up more than, say, 12 hours without rebooting itself. I seem to recall something strange going on with the fans too, but I don't remember enough detail to say for sure. Only a complete, pull-the-plugs-from-the-wall power cycle resolved it.

Worse was that I was trying to resilver two replacement disks into the pool at the time, and that resilver operation was interrupted several times by spontaneous reboots. That ultimately completed successfully, but somehow in the process the device identifiers changed, so I have this:

I'd posted a bug at the time (on Redmine, I think; this was probably before Jira's time) and they were able to help me figure out the connection. Weird thing though.