6.5 few issues

[raymix]

Hi, folks.

Update assistant log: https://pastebin.com/b2Wgsjrk

 

#1 - EDIT: preclear not supported since 6.4.1

I'll start with the smallest one - preclear addon keeps dissapearing on reboot after the update. In fact, I think I have mysteriously lost a ton of plugins, but majority of them I never installed. Any ideas? It seems to keep .disk file, but not .plg.

Also, preclear does not appear in CA anymore, I have to manually install plg link from github. It works fine until reboot, basically.

 

#2 - EDIT: bug

Plugins section takes 68 seconds to load with only 11 plugins showing installed.

 

#3

And finally - 6.4.1 introduced UDMA CRC error count. I am upgrading from 6.4, so I never saw these warnings until now.

I am getting hammered with warnings send to my my email every 30 minutes about both of my cache drives. Because of this I had to disable all email alerts.

I have reseated cables, cables are brand new, barely 2 months old, they do not cause any issues on other drives.

Both 2.5" disks are the same, used as cache drives - TOSHIBA MQ01ACF032 in a Icy Dock MB322SP-B adapter dock, that way they share one SATA power connection and makes use of 5.25" bays.

For SATA controller I use cheapo Syba SI-PEX40071 SATA III 8 Port PCI-E 2.0 as there were no other better devices at this price range and port count.

 

Errors on both drives:

Event: unRAID Cache disk SMART health [199] Subject: Warning [UNRAID] – udma crc error count is 92831193 Description: TOSHIBA_MQ01ACF032_748XCZ4UT (sdg) Importance: warning

Event: unRAID Cache 2 SMART health [199] Subject: Warning [UNRAID] – udma crc error count is 84380027 Description: TOSHIBA_MQ01ACF032_748XCZ5GT (sdh) Importance: warning

 

I have been running without any issues since 6.4 was released up until now. Acknowledging issues on unraid webui is, ofcourse, pointless.

 

I have attached diagnostic logs zip below, hopefully forum allows that.

 

 

 

unraid-diagnostics-20180319-2240.zip

 

P.S.

Also, any idea what's the story with Sent mail for [email protected] in syslog? I googled it, and quite a few other people here has it, but nobody seems to care so I am assuming this one's ok?

Edited March 20, 2018 by raymix

[itimpi]

The pre-clear plugin is not compatible with 6.5 which is why it does not appear  in CA

 

getting a small number of crc errors is normal - but getting the number you seem to be getting is not!   I have no idea what could be causing them in such a quantity, but I would have thought they were adversely affecting performance.   You can always stop monitoring that SMART attribute, but it seems a good idea to get to the bott0m of why you get so many in the first place.

[Squid]

32 minutes ago, raymix said:

I'll start with the smallest one - preclear addon keeps dissapearing on reboot after the update. In fact, I think I have mysteriously lost a ton of plugins, but majority of them I never installed. Any ideas? It seems to keep .disk file, but not .plg.

Because it's not compatible.  The Update Assistant would've pointed that out if you ran it before updating.  After updating, because it didn't install, Update Assistant passes.  Not available in CA on 6.5.0 because its incompatible with 6.4.1+

 

34 minutes ago, raymix said:

Plugins section takes 68 seconds to load with only 11 plugins showing installed.

Bug - It's automatically checking for updates.  Will do that anyways if you have unRaid set to check for updates on some schedule, but if you don't then it checking is a bug.

[JorgeB]

41 minutes ago, raymix said:

Both 2.5" disks are the same, used as cache drives - TOSHIBA MQ01ACF032 in a Icy Dock MB322SP-B adapter dock, that way they share one SATA power connection and makes use of 5.25" bays.

Try without the Icy Dock.

[raymix]

Thanks, guys. What alternatives are you using for your preclear needs?

 

8 minutes ago, Squid said:

Bug - It's automatically checking for updates.  Will do that anyways if you have unRaid set to check for updates on some schedule, but if you don't then it checking is a bug.

 

6.4 didn't have update assistant. If it had, I would have probably stayed with 6.4 if I saw it would get removed.

You're right, I have it set to auto update specific plugins daily at 12AM. I have now disabled auto update and schedule for plugins, but it's still taking over a minute. Disabling docker auto updates did not help either, I'll switch those back on again. Looks like a bug, indeed.

 

3 minutes ago, johnnie.black said:

Try without the Icy Dock.

I'll try with second Icy Dock (has no problems) and without it tomorrow after the work, thanks.

 

[trurl]

3 hours ago, raymix said:

I have attached diagnostic logs zip below, hopefully forum allows that.

Not only do we allow it, we greatly prefer it to external links, or incomplete diagnostics such as only a syslog or SMART report. Please always attach complete diagnostics zip.

[Squid]

12 hours ago, raymix said:

6.4 didn't have update assistant. If it had, I would have probably stayed with 6.4 if I saw it would get removed

Its part of Fix Common Problems actually.

[trurl]

13 hours ago, raymix said:

what's the story with Sent mail for [email protected] in syslog?

Anonymizing your email address so it won't appear in your logs when you post them.

[raymix]

Hi guys

 

Thanks for all the replies.

 

I have swapped drives around with drives in second icy dock caddy (no errors there), but I am still getting UDMA errors, they appeared instantly during boot.

Looks like I am getting 300'000 UDMA errors per day on both drives, so going by this logic, this has been happening for almost a year now (84745745 / 300000), I do not notice any particular slowdowns or anything like that.

 

So, if it's not a cable or SATA controller, this means it's probably one of drives that's bad - am I right to assume that since both cache drives run as "RAID1" so CRC errors would show on both drives?

Is there anything else I could do to test them before I toss them out of window and maybe get some SSDs instead?

Actually, do I even need cache drives for plex media server sharing 720p/1080p content?

I have noticed that I can disable "SMART attribute notifications, Attribute=199" individually, that could work, I guess?

[JorgeB]

1 minute ago, raymix said:

am I right to assume that since both cache drives run as "RAID1" so CRC errors would show on both drives?

No, they would appear just on one, it's very unlikely both SSDs would have the same problem, cables, enlosure, controller are the most likely candidates, but it's not impossible.

[pwm]

3 hours ago, raymix said:

I have noticed that I can disable "SMART attribute notifications, Attribute=199" individually, that could work, I guess?

That's like putting black tape over the engine temperature warning light and continue to use the car without coolant in the radiator.

 

You have an issue you need to solve. If you have a USB enclosure you could test to put a drive in that enclosure and see if you still get the error to count up. If it does, then you can be almost certain it's the drive that has the issue. If it doesn't, then you need to suspect the cables or the controller. Or possibly bad power to the drive.

 

The problem with a large count of UDMA CRC is that if you get enough of them, you may get a transfer error that computes a correct checksum. So instead of a retransmission (that loses you a tiny fraction of the bandwidth) you might now and then get broken data to or from the drive. If it's to the drive, then you will permanent that broken data on the disk surface resulting in a permanently broken file.

 

As noted by @johnnie.black, it doesn't matter if you use drives in a mirror or not. The UDMA counter will only tick up for the specific drives that suffers transfer errors - if both drives tick up then both drives have issues with the communication.

[raymix]

Thanks, I'll give USB to SATA cable a go and try to run some tests on it. Any particular software that you guys recommend for tests on Windows?