[Support] binhex - Preclear

[FreeMan]

I'm preclearing a brand new drive (it hasn't finished yet, so no report to post) and I've got "Warning [BACKUP] - reallocated sector ct is 8". That's not a good sign, is it...

[binhex]

1 hour ago, FreeMan said:

"Warning [BACKUP] - reallocated sector ct is 8". That's not a good sign, is it...

that does not bode well, no 😞 i would seriously consider RMA if possible, let if finish though eh.

[FreeMan]

27 minutes ago, binhex said:

that does not bode well, no 😞 i would seriously consider RMA if possible, let if finish though eh.

le sigh...

 

Honestly, though, this is the first drive I've ever had fail during a preclear. I'm glad I ran it!

 

It's currently at 86% of the post-read and chugging along at rock-steady 200.1 MB/s so it's moving nicely. I'll plug it into my Windoze desktop via a USB adapter and run Seatools on it then RMA it.

[binhex]

2 minutes ago, FreeMan said:

le sigh...

 

Honestly, though, this is the first drive I've ever had fail during a preclear. I'm glad I ran it!

 

It's currently at 86% of the post-read and chugging along at rock-steady 200.1 MB/s so it's moving nicely. I'll plug it into my Windoze desktop via a USB adapter and run Seatools on it then RMA it.

ok cool, post the end report if you get chance, i would be interested in seeing the SMART attributes for the drive after a preclear run.

Edited January 5, 2020 by binhex

[FreeMan]

@binhex Here's the preclear with errors as requested.

 

preclear_rpt_Z303DVL4_2020-01-05

 

Interesting to note that all the speeds listed are in the 120-140 MB/s range, yet the display consistently showed 200.1 MB/s during the post-read and was showing in the 170-180 MB/s range for the pre-read. (not just at the beginning of the disk, towards the middle, too.)

Edited January 5, 2020 by FreeMan

[franky52]

Hi!

Everytime this plugin is installed or started, the GUI gets corrupted.

 

I´m testing Unraid 6.8.1, but it happens with other versions too.

 

 

[Frank1940]

Try white-listing the Unraid URL in your browser ad-blacker or privacy extensions you have installed.

[franky52]

Thank you, but that didn´t help. It only happens with this plugin.

[binhex]

3 hours ago, franky52 said:

Everytime this plugin is installed or started, the GUI gets corrupted.

ok things to try:-

 

1. try a different browser - this should rule out any cookie and caching related issues

2. try a different machine/tablet/phone - this will rule out the client device being the issue

3. post your docker run command - its possible the issue is related to a port clash, to get the run command edit the container, make a small change and then revert then click on apply at the bottom, this will then display a window, copy the entire 'command' section and paste it here.

 

[franky52]

Thank you. I solved the problem reinstalling the system and reseting the arraid.

 

I´m testing Unraid and apparently the downloaded docker image was corrupted or incomplete. I was able to start the docker and even browse through the folders but something was missing.

 

The plugin is working now.

 

[binhex]

Thank you. I solved the problem reinstalling the system and reseting the arraid.
 
I´m testing Unraid and apparently the downloaded docker image was corrupted or incomplete. I was able to start the docker and even browse through the folders but something was missing.
 
The plugin is working now.
 

Just for reference this is NOT a plugin, plugins run on the host operating system, this is a docker image and thus runs in a sandboxed environment.

Sent from my CLT-L09 using Tapatalk

[TechMed]

Awesome container @binhex! And, thanks @Frank1940 for the documentation!

I cleared two 10Tb simultaneously (took a bit under 50 hours for a single pass) with the following command: preclear_binhex.sh -f -c 1 -M 4 /dev/sdo …(and sdp); ran great!

 

Two questions:

1.     I am pretty sure I already know the answer but I just want to be sure: I did not receive any emails, even with M 4 being set. I am betting it is because I have “Notices” turned off under Notification Settings, correct?

 

 

 

 

2.     While the reports (I am ECSTATIC that we can go look at the results well after the fact) indicate the drives both “…[have] been successfully precleared,” it does not show that way in UD. The one showing as “Precleared” in the image below is from a previous run with the preclear PLUGIN, not this container. Thoughts? I am not concerned so long as the drive does not clear again when finally put into the array; not a deal breaker by any means though, just an ugh.

 

So, GREAT job on the container, thank you, and when money for more drives permits, I give the mail option another try.

[Frank1940]

Let's try something.   Settings    >>>  Notification Settings   >>>  SMTP Settings   and click on the "TEST" button. You should get an E-mail within a few seconds.  (Within the past couple of weeks, I helped someone who forgot to put in an E-mail recipient!)

[TechMed]

It works. Actually, it is because I was getting 62 emails a day that I trimmed what all sent emails in Notification Settings 😉.

 

Here is the output from the Test:

Event: Unraid Status

Subject: Unraid SMTP Test

Description: Test message received!

Importance: alert

[binhex]

12 hours ago, TechMed said:

it is because I was getting 62 emails a day

not sure what was triggering that spam, but that is not normal, i would HIGHLY advise switching notifications on, knowing when a disk fails SMART or when it has a read or write failure as soon as it happens is IMHO critical info.

[TechMed]

Sorry for the misleading comment; the “62” was just exaggeration for illustration purposes. It was just that I was receiving an annoying amount of emails with the “Notices” via email also turned on. I was just trying to separate the email wheat from the chaff; I have since turned it back on.

 

During my reading on the setup/configuration of the container, I noticed that Dynamix.cfg was being used and I checked it to make sure my smtp email info was in there, and it is. And yes, I did add the two additional Path’s as indicated in the instructions before starting the container and preclear process. I am under the impression that if an email addy is not specified in the preclear cmd line script, that it defaults to the system email, is that correct?

 

Also, this was the output from the “Parted -l” for the two drives:

Model: ATA WDC WD101XXXXXX (scsi)

Disk /dev/sdp: 10.0TB

Sector size (logical/physical): 512B/4096B

Partition Table: msdos

Disk Flags:

 

Number  Start  End  Size  Type  File system  Flags

Is this correct, particularly the “msdos” partition table?

 

Finally, is this process like (way) back in the day when you had to run “fdisk” on a new drive before you could format it?

[binhex]

19 hours ago, TechMed said:

I did not receive any emails, even with M 4 being set. I am betting it is because I have “Notices” turned off under Notification Settings, correct?

correct.

 

19 hours ago, TechMed said:

While the reports (I am ECSTATIC that we can go look at the results well after the fact) indicate the drives both “…[have] been successfully precleared,” it does not show that way in UD.

i would advise against any sort of mount of the precleared drive via UD, as soon as you write ANYTHING to this disk then the disk being 'cleared' is invalidated and will be cleared again by unraid when you add it to the array.

 

The best way to prevent a clear from happening is to unplug the drive as soon as you can after a preclear has been run, this prevents anything being written to the drive and thus triggering the clear.

 

I myself have seen a clear of a drive that i previously had pre-cleared, i put this down to something (plugin most probably) writing to the drive and invalidating the signature. On another drive prior to that (same make and model and size) i did not have this issue, and i had pulled the power on that drive fairly quickly after a preclear (2 hours), so im reasonably confident the preclear script is working as intended.

[binhex]

13 minutes ago, TechMed said:

Is this correct, particularly the “msdos” partition table?

no it shouldnt be showing as msdos, this sounds like a bug i had a long while ago, see link below:-

https://forums.unraid.net/topic/83465-support-binhex-preclear/?do=findComment&comment=776646

 

are you sure you are up to date with this docker image?, youve done a 'check for updates' and downloaded the latest right?.

 

can you paste the command line you used.

 

Quote

I am under the impression that if an email addy is not specified in the preclear cmd line script, that it defaults to the system email, is that correct?

correct.

Edited January 30, 2020 by binhex

[TechMed]

13 minutes ago, binhex said:

no it shouldnt be showing as msdos,

I read this entire thread before starting and it was that post, about it showing up as "msdos" that made me question the result(s).

So... just "format" them? Not sure I understand... sorry 🤔

 

16 minutes ago, binhex said:

are you sure you are up to date with this docker image?, youve done a 'check for updates' and downloaded the latest right?.

Yup! Pretty anal about those sorts of things.

 

16 minutes ago, binhex said:

can you paste the command line you used.

 

 

18 minutes ago, binhex said:

correct.

Cool. Thanks for confirming.

[binhex]

24 minutes ago, TechMed said:

So... just "format" them? Not sure I understand... sorry

im not certain of what would happen if you added these disks to the array, i think you would end up with the disk being shown as a total size of 2TB, as this is the size limit of a msdos partition, i dont think a unraid 'clear' (if it is indeed triggered) would delete and re-create the partition - the safe option is to re-run a preclear again, this time try running it with the additional -A flag specified (should not be required).

 

couple of questions for you:-

1. can you tell me what your setting is for unraid webui/settings/disk settings/default partition format - is it set to MBR unaligned?

2. how is the drive connected to your server? - what connector and adapter (if any) are you using?.

[TechMed]

I am happy to assist.

Just need a few to get some other things wrapped up.

Talk soon...

[TechMed]

Absolutely no rush at all @binhex, I mean that.

When you have the spare time, let me know on the below.

Thanks again for all your hard work!

------------------------------------------------------------------------------------------------

Apologies for the delayed reply...

On 1/30/2020 at 12:12 PM, binhex said:

This time try running it with the additional -A flag specified

Do you want me to use the 10TB drive(s) or something smaller?

 

On 1/30/2020 at 12:12 PM, binhex said:

/default partition format - is it set to MBR unaligned?

Yes. I have attached a screenshot of the Disk Settings anyway.

Disksettings_01-20.pdf

 

On 1/30/2020 at 12:12 PM, binhex said:

how is the drive connected to your server? - what connector and adapter (if any) are you using?

Via Chenbro NR40700 SAS backplane -- SAS connector and cable -- (from Tools >>> SysDevs) IOMMU group 43:[1000:005d] 01:00.0 RAID bus controller: Broadcom / LSI MegaRAID SAS-3 3108 [Invader] (rev 02)

JIC you need to know, this MB is a SM X10DRH-CT  V 1.1  - Dual Xeon E5-2620 V4 - 64GB - a mix of 14 WD and HGST drives - Crucial MX-500 Cache

 

So, let me know if you want me to test the 10TB drive or something smaller. Also, Is this the command you want me to use?::  preclear_binhex.sh -f -c 1 -A -M 4 /dev/sdX

 

 

Edited February 2, 2020 by TechMed
Spelling and ping

[binhex]

On 1/31/2020 at 8:43 PM, TechMed said:

Do you want me to use the 10TB drive(s) or something smaller?

ok lets try something smaller for now (just not smaller than 2TB), this may give us an indication as to whether its UD causing the issue (see comments below for theories).

 

On 1/31/2020 at 8:43 PM, TechMed said:

Also, Is this the command you want me to use?::  preclear_binhex.sh -f -c 1 -A -M 4 /dev/sdX

yes thats it.

 

so at the moment i am considering two possibilities as to why it has created a 2TB partition on your 10TB drive:-

 

1. Unassigned Devices is interfering with the drive during preclear - this is my top suspect at this time, although its obviously difficult to rule out without stopping/disabling the plugin, which you probably dont want to do.
 

maybe what you could try is just try and stay clear of using the unassigned devices web ui if possible, at least whilst the drive is in the process of being precleared.

 

2. bug in the script that affects drives over a certain size (not determined yet) - so this is a possibility, but a little unlikely, i have used this image to preclear two 8TB drives without issue, so it def can handle drives over the 2TB mark, but i guess it is possibly that it has issues with drives larger than 8TB perhaps, this is of course tricky for me to test, as i dont have a spare 10TB+ drive just knocking about, but you do :-).

 

 

[TechMed]

Edit: 02/11/20

@binhex, just checking to see if you'll have any time this week to mess with this. No worrries if not, just checking in. Later...

On 2/5/2020 at 8:34 AM, binhex said:

i dont have a spare 10TB+ drive just knocking about, but you do 🙂

So yeah, I stumbled across them as factory refurb’s, < $200 each, direct from WD, and couldn’t pass them up. Caught six kinds of H@(( for spending the money, but… Plus, I do not have the overhead anymore like you do. Which reminds me, are ‘congrats’ in order yet? ☺️

 

Okay, I ran the 10TB again anyway, with the command discussed (adding the -A) and it just straight up disappeared at or around the end of the ‘zeroing’ phase. Disappeared meaning, I checked on it (via No VNC) before bed and it was at 86% ‘zeroing’ then I shutdown my desktop. Started up desktop next day expecting to be able to see the end results like before, but there was nothing! No logs, no errors, no nothing. It was as if I never even ran the disk through preclear. That was the morning of February 4. So, I collected diagnostics, drive logs, preclear logs and have them all zipped up and attached. Note: some of the logs include old data from the Preclear Plugin.

 

On 2/5/2020 at 8:34 AM, binhex said:

Unassigned Devices is interfering with the drive during preclear - this is my top suspect at this time

I am not good enough yet to understand all of the error messages, but there appear to be a a number of them related to UD at the bottom of the “Disk Log Information” log.

****EDIT: It just dawned on me that I think there was an UD update that night. If UD updated, that would have hosed the preclear, wouldn't have it???***

On 2/5/2020 at 8:34 AM, binhex said:

try and stay clear of using the unassigned devices web ui

Sorry, not sure I understand. Are you saying here to uninstall UD then run Preclear?

 

Also, I am still not getting any notifications at all. I will try adding my email addy next go round but, Frank1940 had a good idea and asked me to test my SMTP settings, but they worked fine. Dunno?

 

Let me know what you want me to try next and I will get it going using the 10TB. I will post preclear progress when I can; busy end of the week with better availability over the weekend.

Diag Zip Files.zip

Edited February 11, 2020 by TechMed
Epiphanic moment ;) - ping

[FreeMan]

Very odd...

 

I picked up a Seagate Expansion 8TB on sale from NewEgg for $125 (not a bad deal) and it came with a green-labeled Barracuda drive, so I'm happy with my purchase!

 

Using this docker, I ran a preclear (about 51 hrs) with the drive still in its external enclosure. (I pointed a small desktop fan at the vents at the back and had the case sitting with the bottom vents on top - never topped 35°C!) This morning, after finishing the preclear late last night, I unplugged the USB cables, powered the drive off & shut down my server. I shucked the drive (BTW- I don't think Seagate wants you to do that anymore. I ended up using a metal putty knife to pry the case open - plastic ones wouldn't do it - and did enough damage that I doubt they'd take it back under warranty) and installed it in my server. 

 

When I booted the server up, I stopped the array, added the drive, started the array and it's now 26 minutes and 4% into clearing the drive before adding it to the array.

 

I've attached the preclear report - everything there indicates to me that it completed successfully. Any indication of why UNRAID may have thought that the drive was not cleared? I did look at the list of devices in UD, just to make sure it had been detected, but I did not mount the drive - I know better than that... As far as I know, nothing would have written to it overnight.

preclear_rpt_ZCT1199D_2020-02-21