xrqp

Members
  • Posts

    200
  • Joined

  • Last visited

Everything posted by xrqp

  1. I am not sure which vendor you refer to. Maybe Amazon vendors "tech on tech" and "hyperhawk"? Even though there were high % DOA, they all checked out as having full warranty. I don't know if high failure rates is the vendors fault since the shipper and the manufacturer are also involved.
  2. Good to hear that. That is ideal. Can it also be updated from the Unraid docker screen? Does it do the same thing either way?
  3. I ran a 1st correcting parity check and it reported 256 errors, and I think it fixed the errors, because the next (2nd) parity check without error correction showed zero errors. So should I relax and go back to normal use of unraid? 1st tower-diagnostics-20220503-1728.zip2nd tower-diagnostics-20220505-1700.zip
  4. Thanks for replies. I will try updating the core and all, from the nag next time, and pay attention if it works. I think for most programs we should only update dockers from the unraid docker tab, but the Roon core docker is unusual, so I guess it does not apply.
  5. How to handle Roon updates if the Roon core is on Unraid NAS in a docker. Trying to refresh my memory on this: Is it correct that: We should only upgrade the Roon core via Unraid's docker page? We should NOT update the Roon core via Roon app nagging to update? If you updated the core from the Roon app, it does not work, but it does not harm anything? When the Roon app gives a nag to update, it gives a choice of "update all" or "view details" or "x" to close the nag. I clicked view details then it showed that it is my Unraid Tower that it wants to update. Here is screenshot: I think one time i clicked it to "update all", but nothing happened to the core. So it seems the core update fails, but maybe there is no harm done to the core.
  6. I will run correcting parity check, get diagnostics, then run non correcting parity check and get diagnostics. Each parity check takes about 2 days. One problem I have is my internet is DSL on a very noisey copper wire phone line. I have voice phone (POTS) on the same line and it is so noisy we can not use it for voice anymore.
  7. I ran memtest for about 24 hours. When I stopped it, it said "Pass:10 Errors:0". It was running test #9 when I stopped it, so I do not know what "Pass:10" means. I think I passed the memtest. So I am now running unraid again. So now what do I do?
  8. No, I was going to keep running it until Monday. But based on your reply, I will start memtest as soon as I get home tonight.
  9. Thanks for the link. You have confirmed my hunch. The link said If the hard drive has firmware locked, users cannot reset smart at all. Users must use DFL firmware repair tools to unlock the SA and then users can click reset smart to complete the process. I assume that first sentence was just not written well to say "not at all". So from the 2nd sentence it appears they can reset smart data in all cases. Depressing. So the only protection is to check warranty. If the 5 yr warranty ends in 5 years, then the drive is probably new.
  10. I read that I can run memtest from the boot menu. To see boot menu, i hook up monitor and keyboard then reboot. I will try to run memtest next week. I tried to figure out if I have UEFI or not, by looking for folder on flash drive for efi~ or efi, so I can make sure it is efi~ . but when I use Krusader, and go to Root/FLASH it says I cannot open that folder.
  11. Should I run a correcting parity check? It is hard to do parity check without corrections as there are 2 places it is set to correct. In the scheduler I changed it to not correct, but in the main window, near the bottom, next to Parity check, there is a little check box "correcting". I uncheck it, but when the screen refreshes, the check comes back by itself.
  12. I will see if I can put a watt meter on and stress the workload, and see if I exceed my power supply, which i think is 400 watt (need to check if 350 or 400 watt PS). When I had 5 drives it peaked at about 90 watts and idled at 35 watts, but I did not try hard to max out CPU.
  13. On 4/25 email, I did think Health check was the same as parity check. I realize now they are different. Thanks. I will check forum for how to do memtest in unraid. Could errors be caused by too much going on at one time? I may have been transferring files from Wind 10 PC to unraid, and downloading tv shows from antenna, and downloading from sonarr and sabnzbd thru web, and running roon docker. Should I run a correcting parity check? It is hard to do parity check without corrections as there are 2 places it is set to correct. In the scheduler I changed it to not correct, but in the main window, near the bottom, next to Parity check, there is a little check box "correcting". I uncheck it, but when the screen refreshes, the check comes back by itself.
  14. At this moment April 26 after the last parity check, on main page the column for "errors" all are zeros, the dashboars shows "healthy" for all disks, and I attached a new diagnostics file. tower-diagnostics-20220426-1935.zip When i wrote on April 25th that I could not find errors, it was based on checking syslog, and also checking SMART report for every disk. Based on emails below, it looks like this happened over the week: Parity check - result 4/21 - 278 errors 4/25 - pass 4/26 - 256 errors. I was mixed-up before when I started this thread - I thought the 278 errors were on 4/25. Here is the last 2 emails I got when I ran parity checks. April 25 ran w/correction. April 26, ran w/o correction. The April 25 email included info on the previous parity check on April 21 with 278 errors. EMAIL Sent: Monday, April 25, 2022 12:20 AM Subject: Unraid Status: Notice [TOWER] - array health report [PASS] Event: Unraid Status Subject: Notice [TOWER] - array health report [PASS] Description: Array has 9 disks (including parity & cache) Importance: normal Parity - ST18000NM000J-2TV103_ZR52HS8A (sdg) - active 40 C [OK] Disk 1 - ST18000NM000J-2TV103_ZR52VKR5 (sdk) - active 37 C [OK] Disk 2 - ST18000NM000J-2TV103_ZR52BE8Q (sdi) - standby [OK] Disk 3 - ST8000DM004-2CX188_ZCT06E5C (sdd) - standby [OK] Disk 4 - ST8000DM004-2CX188_WCT0VFC8 (sdh) - active 31 C [OK] Disk 5 - ST8000DM004-2CX188_ZR119PYR (sde) - active 30 C [OK] Disk 6 - ST18000NM000J-2TV103_ZR52HRY7 (sdf) - standby [OK] Disk 7 - WDC_WD80EFAX-68LHPN0_7HKBDRTF (sdj) - standby [OK] Cache - CT1000P1SSD8_1842E1D21EFE (nvme0n1) - active 38 C [OK] Parity is valid Last checked on Thu 21 Apr 2022 10:50:21 AM PDT (4 days ago), finding 278 errors. Duration: 1 day, 10 hours, 50 minutes, 20 seconds. Average speed: 143.5 MB/s EMAIL Sent: Tuesday, April 26, 2022 1:30 PM Subject: Unraid Status: Notice [TOWER] - Parity check finished (256 errors) Importance: High Event: Unraid Parity check Subject: Notice [TOWER] - Parity check finished (256 errors) Description: Duration: 1 day, 12 hours, 27 minutes, 50 seconds. Average speed: 137.1 MB/s Importance: warning
  15. OK. I am running parity check with no corrections now. Takes about 2 days. I did not correct any errors because I could not find any. thanks again.
  16. I numbered your lines to reply to each. 1. I do not think we can know the reason for failures, if it is bad production, shipping mishandling, warehouse conditions etc. At least I did not research that. 2. On Amazon, all in 2021, i bought from "tech on tech" once (and had a exchange w 2nd one good), and "hyperhawk" 2 or 3 times and had exchange w 2nd one good, about 2 times. I also bought from non-Amazon as described above in the thread, and I think it was no return problems from non-amazon, I can't remember for sure. 3. Low price. I know I can test it when I get it, and I can check warranty, and if problems, it is fairly easy for me to return. UPS store is very nearby. 4. As soon as I receive, I always go online, enter the serial and model and check warranty, regardless of vendor. So far I have not had any problem with that for the Seagate 18TBs. I also do Crystal Disk Info to make sure it is unused. I also do unraid preclear once or twice. No warranty problems, no preclear problems. Only DOA problems so far. I think a vendor who sells a lot, could trick Crystal Disk Info by resetting some of the hard drive metadata (maybe it is in firmware or a protected area of the disk). Do you think that there is a good chance they can do that?
  17. So syslog is the place I can see the errors listed? I got syslog by going to "Tools" then "syslog". Is syslog also in diagnostic.zip? I can not find it in diaglnostic.zip. In syslog it says: I had about 30 red errors for "no such file or directory". Those do not worry me. Not like a disk error. Those are porobably because I setup Sonarr imperfectly. I also had about 30 green logs saying "Tower kernel: md: recovery thread: P corrected, sector=4015656792". Each error with a different sector number. Based on your reply, I clicked on "scheduler", then changed "Write corrections to parity disk" from yes to no. (Unraid should give better help on that one). I do auto parity checks every 3 months. I guess I am ok then, and can just wait until the next auto parity check in 3 months. Thanks for your reply and help.
  18. Should I just ignore the parity check saying there are errors? tower-diagnostics-20220424-1703.zip I saw the little warning box saying I had 300 errors (rounding off) after parity check. Is there a log or file I can see those errors listed? Where?
  19. Newegg has user reviews of the 18TB Exos, and I interpret it as one "Dead on Arrival" for every 5 ordered. My experience is worse, but I still ordered one more today for $310 from serverpartdeals.com w free ship to USA. before, my usual source was Amazon, but they have none today. I am depending on the 5yr warranty and hoping if not DOA it will be OK. If DOA I return it. Of course I'd rather that the DOA rate be much much lower. Maybe the 18TB HGST, for $350 from serverpartdeals.com are a better choice? I don't see much info about the DOA rates, but the low quantity info shows them to be much better in that regards, but only 17 reviews at newegg, so not too meaningful (1 DOA per 17).
  20. I do not understand your reply. I meant the physical disk will not be affected, so will keep its preclear. I am not concerned about separate things like the log. I assume the log is not stored on the subject disk, and that the preclear signature is stored on the subject disk.
  21. If I click the red x, does the disk keeps its precleared state and remain precleared?
  22. I am not sure if I am using the correct pre-clear. I use the one by clicking the paintbrush on the right side of this graphic: Does that mean I am using the "unassigned devices pre-clear" which is OK, and that I am NOT using the problematic one by gfjardim? UPDATE: I removed the gfjardim preclear plugin, and now the paintbrush is gone, so now I need to figure out how to do the unassigned devices pre-clear. UPDATE: For idiots like me, that have not done a plug in for so long I forgot how: Click on "Apps", then on the text list on the left side click on "plugins", then in the search box near the top type "unassigned" or "unass" and it will be easy to find "unassigned devices pre-clear". After installing the plugin, on main page, you will see, if you look carefully, a very small paint brush (in monchrome blue) on the right as shown on this graphic: I was looking for the big colorful paintbrush, so it took me awhile to notice this. I did not need to reboot to see it. Thanks to dlandon for the plugin. Info at https://forums.unraid.net/topic/120567-unassigned-devices-preclear-a-utility-to-preclear-disks-before-adding-them-to-the-array/?tab=comments#comment-1101719 For the sake of completeness on this set of graphics I am posting here (otherwise I would post this at the link above). I clicked the blue paintbrush and selected "Verify Signature" (as I had done gfjardim preclear months ago) and in 20 seconds it verified the disk was already in a good preclear state. And the graphic changed to this: then I clicked on the black paintbrush which did nothing and clicked on the "PRECLEAR" gray button which did nothing. So I could not run the "Verify Disk" (where the disk signature is checked and the disk is verified to be zeroed).
  23. I think maybe a few files got messed up while being used by an app, or some kind of glitch. I was able to delete the files using Krusader and all seems to be OK again.
  24. It was a public share at the time of the problem. So I tried secure share and same problem.