HSEDuster

Short answer, no. Long answer, no, but after a lot of searching, I think it might be related to something to do with either the SAS drives or the HBA. I only guess that because I can find some (limited) information on ASC and ASCQ "codes".. It's pretty lame though as these eventually fill up my log file and as far as I know the only way to clear that is a restart... Out of curiosity, what's your setup? Just wondering if there's any common denominator here...

This is actually super helpful. I feel like an idiot for not thinking of this. I access my unraid servers at work sometimes when things are slow (err like now), and never thought of this to get into my dockers...

For a few months now my primary unraid server experiences seemingly random times when it becomes unresponsive. That is, SMB and NFS shares are inaccessible, VM's don't work, and I can not access the web GUI. I don't know for sure, but I believe it might be connected to a repeating log entry I keep getting: Jan 31 16:39:54 Tower kernel: sd 3:0:6:0: [sdh] tag#1347 Sense Key : 0x1 [current] Jan 31 16:39:54 Tower kernel: sd 3:0:6:0: [sdh] tag#1347 ASC=0x1c ASCQ=0x0 That's just from one disk, but all the disks get them and its absolutely unrelenting. Sometimes the server will run for weeks, though more often it will run for a few days before needing to be hard-reset. While I can't recall exactly, I know this started happening after I moved my server from an R410 to an R710. All of my drives are (and have been) SAS drives. After the upgrade, however, I did replace my SATA cache SSDs with LB806M SAS SSDs. Those are mounted in the actual R710 with the HDDs being mounted in an MD1000 external case. Diagnostics attached, and I would be happy to provide any other information needed. Side note, I don't know what it will show in diagnostics, but when downloaded, I was (am) running a parity check with Disk 3 missing. The drive was failing so I replaced a parity drive with a new 10Tb and once this parity check is done, the old 8Tb parity will go in Disk 3's place. There is also a UD that is 'unhealthy', but it's an old desktop drive I was pulling files off of, nothing related to the system as a whole. Just figured I would mention it. tower-diagnostics-20220131-1638.zip

You are totally correct. So much for my 30+ day uptime streak though... Ah but I'll take the hit. Thank you very much!

Hey guys. Short version is I had two 800GB SSDs in a pool, and last night added two additional 800GB SSDs, each in their own respective pools (FWIW the original is BTRFS with the two new ones each being XFS). I happen to notice that *all* my shares are missing. The data is fine, just the shares are missing. Dockers seem to run fine, though I haven't tried all of them, and one of my VMs worked after I re-mapped the location for the vdisk on one of the new SSDs. Also might be relevant, I renamed all three of my cache pools. Diagnostics attached. Side note, I'm fairly sure there's a few more issues that might be noticed in my diagnostics, there are a few things I'm working on, but this issue jumped to the head of the line since I can't get to any files (at least not easily). tower-diagnostics-20210525-0842.zip

I don't know if its "optimal" but I have an R710 with an H200 for the front drives and an LSI card for drives in an external DAS. I personally have my array drives, and most unassigned drives, in the external DAS and I have 2 SAS SSD cache drives in the 710 itself, along with 4 470GB unassigned SSDs I have yet to use for anything (my old cache drives). My DAS is SAS 1, that is 3Gbps. The H200 is SAS 2 (6Gbps). While I haven't tried, I'm pretty sure if I put my cache drives in my DAS, that would slow things down. In your case though, if both cards are the same generation, I don't think it would matter too much. I mean, a spinning drive, or even a few, aren't going to saturate a SAS link. A bunch of SSDs might, but only in heavy use situations. FWIW there is a docker you can get to benchmark your drives and adapter cards. If it's not a huge PITA you could swap things around, do some benchmarks and/or transfer tests and see what works best?

I don't know how much insight I can offer, but I can say I am certainly in the same boat. Like... Almost exactly, down to the hardware. That said, I am leaning towards the PowerEdge route personally, but rather than going on some long explanation, I'll just give a brief reason (plus maybe bringing this back to the top will stir up some more valuable replies!): I currently run an R410 with an MD1000 for my disks. I run my cache SSDs in the R410 and array (and unassigned) drives in the DAS. The R410 has been awesome, but the fact I can only install ONE PCIE card is a buzzkill. I use it for my HBA and that's it. Done. No more. No 10GBe, no USB cards, no PCIE SSD, and no video cards. You say you want to use this for VMs, including a gaming VM. I run a few VMs on my much older 410 (albeit no gaming) along with around a dozen dockers and it it runs just fine... Just don't ask Jellyfin to transcode more than a few streams at a time... I think the 430 can support, what, two PCIE slots I think? You might be good in that area (available slots), though making sure things are good with power draw through the slot and/or the chassis has the required power supply connector (if required on your GPU). That's basically it IMO. A traditional build system will likely have more slots available (depends on motherboard), but you can always get a new board, a bigger case, etc... With the server, you're stuck with the board you got (and the CPUs that go with it) and you're also stuck with the case (and the limitations within). I'm going to shoot for an R730 for the additional room, but if the PCIE slots won't be limiting you, then I would feel pretty comfortable going that route. On the other hand, if you think you might ever (well, within the life of the system...) want to have more than two expansion cards, then the AMD would probably be the way to go. Also, if you want more flexibility. ...Of course you could always do both. A mid-range build for your Ryzen and the leftovers on an 11th gen poweredge for the NAS/Pihole/SAB/etc, leaving the VM duties to the Ryzen. I know I picked up two 410s for like $120 on ebay, so that might not be a too outlandish of an idea.

I apologize and certainly see what you're saying. I will be proactive in the future with the diagnostics files. I really appreciate the help and support on the forums in aggregate. This is the the first time I think I have ever posted an issue. 99% of the time I can solve any issue in looking through these forums and spaceinvaderone videos... Ah... I think I might know the issue then. I was having some other issues a while back and in the process of troubleshooting I replaced the cable between my server and the SAS enclosure. When I worked out that it wasn't the cable, or the controller modules in the drive enclosure (an MD1000), I left both cables attached. I assume that this could be the cause. I'll try unplugging one and restarting and see what comes up. I have new processors to install as well, so I might install those while its down this weekend... Of course I guess the alternative culprit could be the HBA card, which is also due to be replaced soon, so hopefully one of the two fixes things. Thank you guys for the help. I really can't convey how appreciated it is. This place is a gold mine of information.

Oh okay, here you go then. Thank you very much. tower-diagnostics-20201005-1609.zip

So I feel pretty dumb asking this, but which? I downloaded the diagnostics but I somewhat doubt you need/want all the various files. I am including the syslog but if there's another, I'll happily post it. syslog.txt

I have a quick (hopefully) question/issue. I'm sure I'm just missing something, but UD shows my array devices as... well... Unassigned devices... Screen shots included, and hopefully self explanatory. The sd* identification is different between the two, but as you can see the HDD ID is the same.

My array is a lot smaller, capacity wise, than yours, but I use an MD1000 I picked up off ebay and an LSI 9201-16e. I connect it and the drive enclosure with a SAS cable (SFF-8088 to 8064). The setup works wonderfully and I have 15 hot swap 3.5" bays with redundant power supplies and EMM's. Unraid recognized the drives right away (after I flashed the card). Plus, I only use one of the four ports on the card for the enclosure. Oh and the card was 20 bucks shipped and the enclosure was 180 shipped with drive caddies, two power supplies and two EMMs plus a front bezel with key (hey the little things count).

I wasn't really sure if this should be in general or security or docker... It kind of spans all that. The TL:DR version is I have an ISP assigned static IP, a domain from namecheap, and an SSL cert from namecheap and want to access my NextCloud from aforementioned (sub)domain with SSL. I know, and currently, have, a free cert through unraid/ let's encrypt but it's a minor annoyance that browsers throw up a hissy fit saying the site isn't secure. Not an issue for me, but I do have some family members sharing this too and avoiding the security threat warnings would save some explaining for me. I can find endless threads about free SSL certs through let's encrypt (and others), IP tracking (duckdns), reverse proxies, etc etc... But I haven't been able to see a thing about having a static IP and getting an SSL cert via a CSR for this. Now granted, I think the static IP bit is probably a non-issue for these purposes... I haven't set up a reverse proxy (at least not yet...), or anything else really beyond getting a functional nextcloud docker (sorry, 'container') running and the subdomain pointing to the correct IP and port. FWIW I use OpenVPN to access any services I need to if I'm not on my local net (SABnzb, Sonarr, etc...) which works wonderfully. The SSL cert is a paid one from Namecheap (it was a package deal when I got the domain) and if I can get it to work, great, but if not so be it. But alas after weeks (admittedly off and on) of scouring the forums and other sources online, I can't piece together the most straightforward way to do this. Any suggestions or just a nudge in the right direction would be awesome. Oh and if it's of any relevance, at the moment I'm using a consumer (Linksys) router connected to my ISPs line. No modem or anything on my end. I'm apparently part of their Motorola canopy network pending installation of fiber (to be installed by August... of 2018... yeah...) so my indoor 'hardware' on the ISP end consists of a POE injector and a CAT5 cable...