October 8, 20169 yr New Unraid install. I successfully migrated all my data to a new array made up to two 6TB drives. One parity, one data. Both drives are connected to internal SATA ports. All was well. Now I'm bringing in my older 4TB and 3TB drives attached via an external esata enclosure (Probox) via an ASM 1061(1062?) sata card. This is all the same hardware present in my previous Windows (FlexRaid) build. The drives completed the preclear successfully and were added to the array with no problems. Functioned fine for the first two days. Parity check ran last night with tons of errors and then one of the drives was disabled. I assumed that overnight, the drives were put to sleep and the Probox incorrectly turned off during or prior to parity and caused all the errors. So, I followed the instructions to add a disabled drive back in and now the OTHER drive throws a bunch of errors. Both drives passed smart tests and have been in service with no issues in this enclosure. I can only assume I'm having some kind of connectivity issue due to the 1061 or the box itself and compatibility with Unraid. Luckily, there's no data on the two external drives yet. Unfortunately, my backup on the 6tb is not yet complete so all of this is somewhat scary. Compounding the issue is that the log keeps filling up preventing me from reading it: Allowed memory size of 134217728 bytes exhausted. A reboot clears it but then when I try to run parity it starts throwing errors and the log fills up again. At this point, I'm not sure where to start. Ideally, I can get these two drives back out of the array until I can isolate and correct the connectivity issue and keep my data far far away. Any advice?
October 10, 20169 yr Author You bet. Also, I have some updates that might clarify where things are at for you experts. - This morning I got this notification: Notice [JUPITER] - array health report [PASS] Array has 3 disks (including parity) - I have no clue what's going on but Disk 2 reports as "Not installed" however when you click the little file explorer icon, that appears to be where all the new data is being written. Does that mean it's being emulated? - Disk 3 (also an external disk) claims it's fine but I don't believe it based on the failed parity scan and 8 bajillion errors. - Disk 2 does appear in the unassigned devices. - For those of you who are thinking "what an idiot, why would he not have a good backup." the backup is mostly done now and I wouldn't have a full breakdown if things were lost...but it would still suck. Thanks for your help. jupiter-diagnostics-20161010-0837.zip
October 10, 20169 yr Community Expert Why are you using an eSATA enclosure (with its port multiplier) when you have so few drives?
October 10, 20169 yr Author This is a staged transfer and rebuild. I stood up Unraid on a different machine and transferred data over to the two new big drives. Now I'm adding all the drives from the previous array and I'm out of internal sata ports so the drives go to the enclosure. I picked this sata card (ASM 1061) specifically because I had read it worked well with Unraid/linux. I'm not particularly concerned with blazing fast performance. This basically serves media to one or two clients at a time. The primary goals were reliability and safety moving from a Windows Flexraid install. EDIT: Also, I realize the sata card has internal ports too. Also constrained by current power supply so another reason for the enclosure.
October 10, 20169 yr Community Expert I wasn't questioning the SATA card. It's just that the eSATA enclosure is splitting a single SATA port into 4 (port multiplier) which will greatly affect performance when more than one of those drives is accessed at the same time, which is what always happens during parity sync, parity check, disk rebuild. Still not clear why you have no room in the main case since there aren't many drives. Is this intended to just be a temporary solution? If not a different build might save you a lot of time and trouble.
October 10, 20169 yr Author Do you think the enclosure is the root cause of my issue? I was under the impression this was a supported configuration, just not an ideal one from a performance standpoint. Are you saying Unraid just won't work using an external enclosure? Did the use of the enclosure cause the disk to drop out of the array? What does "not installed" indicate in this scenario?
October 10, 20169 yr Community Expert Not saying it can't work at all. Also not saying it isn't the source of your problem. If you start the array without a disk assigned to a slot, possibly because it wasn't detected at the time, then it will forget about it and nothing will be installed in that slot. Then if it is detected it shows up in unassigned devices because it is not assigned. If you assign the disk does it say it will rebuild?
October 10, 20169 yr Author That's a good question. At this point, I'm wondering what the best procedure is. The array is up right now. I hesitate to mess with it until I have a better understanding of what went wrong and what my actions might do. Should I attempt to add the drive again and bring up the array? Would the next logical step be to run parity? When I attempted that before, I got all errors again and the log filled up. OR, should I do a new config and drop the two external drives. Build parity on the two internal drives and then attempt to add the external drives from scratch? I just don't know the product well enough to determine what the results of my actions will be.
October 10, 20169 yr Community Expert If there is not data on the external drives then I would start with this: ...should I do a new config and drop the two external drives. Build parity on the two internal drives... then after parity sync do a non-correcting parity check. It should come back with no errors. This will be a good test of the rest of your hardware. Tell us about the rest of your hardware.
October 11, 20169 yr Author Sorry, new to this. Can you explain what each of those options actually do? I might have sounded like I knew what I was talking about but I was really just parroting things I'd previously read on the topic. Rest of the hardware is pretty basic: Dell i5 8 GB Ram AS 1061 Probox external enclosure 2x6TB (internal) 1x500GB SSD (internal, not messing with that yet) 1x4TB (external) 1x3TB (external) Eventually might add back in: 1x 3TB (external) 2x1.5TB (eternal) So, the steps (if there is no important data on the 4TB and 3TB) would be: New config. Which options to chose? Add 6tb parity back to parity spot? Add 6tb data back to disk 1? Rebuild? Or... Add 4tb data back to disk 2 Add 3tb data back to disk 3 And then rebuild? I don't know what a non-correcting parity check does.
October 11, 20169 yr Community Expert So, the steps (if there is no important data on the 4TB and 3TB) would be: New config. Which options to chose? Add 6tb parity back to parity spot? Add 6tb data back to disk 1? Rebuild? Or... Add 4tb data back to disk 2 Add 3tb data back to disk 3 And then rebuild? I don't know what a non-correcting parity check does. When you say Dell i5, what exactly is that referring to? Is that a prebuilt computer? Got any more details about the specific model? You could New Config with any and all internal disks. When you Start unRAID don't check the parity is valid (trust) checkbox so it will do a parity sync (build parity). Then after it completes you could do parity check. Doesn't matter really in this scenario whether you let it correct parity or not. If you get a lot of parity errors (should be zero) then you can just stop it before it finishes since that would mean something is wrong.
October 11, 20169 yr Author So on that new config screen, there are some options for the type. It's a scary page. Which option should I use? As for specs, here's what Unraid shows under info: M/B: Dell Inc. - 033FF6 CPU: Intel® Core™ i5 CPU 650 @ 3.20GHz But yes, pre-built Dell from a few years back.
October 12, 20169 yr Community Expert So on that new config screen, there are some options for the type. It's a scary page. Which option should I use? As for specs, here's what Unraid shows under info: M/B: Dell Inc. - 033FF6 CPU: Intel® Core™ i5 CPU 650 @ 3.20GHz But yes, pre-built Dell from a few years back. Post a screenshot of the scary page. I haven't seen it lately and am not going there. I was more interested in the specific model of the complete Dell computer, so I could find out more about the case, power supply, etc. online. I guess you aren't interested in doing a custom build, but do you really need more drives than will fit in the case? You can get some pretty high capacity drives these days.
October 12, 20169 yr Author I'm posting a new diagnostics file in hopes that someone can interpret what's going on. Rather than do the new config, I decided to just stop the array and add the dropped disk back in and see what happened. Parity stopped claiming that it was aborted by the user but I didn't stop it. Disk 2 shows a red X but still shows as installed. Disk 3 shows as green but was full of errors. jupiter-diagnostics-20161012-0759.zip
October 12, 20169 yr Author Now I'm getting somewhere. With entries in the log this time I have something to go off of. It's very likely that this is a Linux + Port Multiplier + ASM1061/2 issue. Looks like others have reported strange lock-ups and other errors similar to what I'm experiencing. Bummer because up until those issues for the first few hours the performance was actually pretty solid. So now my job will be to figure out if there is some kind of updated driver or perhaps firmware for the ASM1061. I also have my old Sil3132 available but I had issues with it in the past not passing smart data to the OS. In the meantime, I assume I should follow the process of shrinking the array/removing disks? That way I can get back to using the known good config with internal drives. To confirm, that's New Config->Don't retain configuration? Then just place the Parity and Disk 1 Data drives in the proper slots and rebuild the array? Attached screenshot.
Archived
This topic is now archived and is closed to further replies.