[Behavior when external enclosure fails]

8 minutes ago, trurl said:

^this

Awesome. That makes sense, and it seems like potentially an acceptable risk. The enclosure will be connected to a UPS, so my primary concern is a faulty controller or powersupply. I'll need to rethink this whole setup in the future. Unless anyone thinks this is data suicide I'll proceed (as long as I can get per-disk S.M.A.R.T. reporting via eSATA that is!)

[Behavior when external enclosure fails]

16 hours ago, trurl said:

Do you have a specific enclosure in mind?

Currently I have a Mediasonic HFR7-SU3S2

 

My eSATA card just arrived today. I'm not super worried about performance, although I believe the enclosure and the eSATA card support port multiplication - my main concern is stability and accurate per-disk S.M.A.R.T. which I'm not sure I'll get with eSATA. The enclosure might get returned in that case and I'd pick up something that this forum recommends instead.

 

13 hours ago, johnnie.black said:

If Unraid loses contact with multiple disks it will disable as many disks as there are parity drives, if more disks than parity drives drop out at same time Unraid won't be able to access or emulated them, but a reboot after reconnecting the disks will bring them all online, you'll then need to rebuild the disabled disk(s).

Cool, this is what I was hoping would happen. Does the array stop when this happens? My ideal situation would be:

- Enclosure fails

- unRAID detects this and shuts the entire array down

- As long as I can re-introduce the drives to the array _somehow_ data loss is relegated to whatever was being written at the time

 

Does rebuilding the disabled disks just involve checking the consistency of the filesystem? Or is it a full parity rebuild against whatever data is available across all the disks?

[Behavior when external enclosure fails]

I'm in the process of adding a cheap eSATA enclosure to my unRAID server. I realize this significantly increases the risk of dataloss, since the enclosure itself could easily fail / lose power / get struck by a meteor - but I'm trying to gauge the actual risk to data loss.

 

I understand if two disks fail in my array, I will lose data across both disks, but what does unRAID do if the array loses multiple disks (up to 4 in this case) and the disks are recoverable? Is any data being written at the time lost but the array recoverable? Does the array immediately stop?

 

I'm willing to accept the downtime / risk of losing whatever data was being written at the time if the array is recoverable - but if the power supply to the enclosure fails and I stand to lose four disks worth of data, I may have to find another solution (or just run a single disk in the enclosure).

 

Basically: I know this is a bad idea, I'm just trying to gauge how bad of an idea it is!

[[Support] binhex - DelugeVPN]

Thanks for the response, I'll check with my provider. I had previously asked them if they restrict any ports, but I can see now that they may have assumed I meant outbound ports. I take it that no additional configuration is needed if my provider isn't restricting incoming ports, other than making sure the Deluge incoming port matches whatever port/port range my provider supports?

 

Thanks!

[[Support] binhex - DelugeVPN]

I hope this hasn't been answered a million times already, but apparently the incoming port for Deluge is blocked by default (or at least it is if one uses a custom VPN, like I am). Is there a work-around for this that anyone has discovered? I've been digging through the config to try to make heads or tails of how I might make the appropriate changes to iptables myself but I'm a total novice.