Last night, i noticed that i couldn't get onto plex.  Tried logging into the server via a windows machine and coudn't bring up the dashboard.  

Its happened in the past but hasn't in a long time.  

I went ahead and did a hard reset of the server and everything was fine last night.  

 

I woke up this morning and Plex was unresponsive again.  I tried logging into the server and same thing.  Non responsive.  

I didn't really know what to do, so just gave the server a hard restart again....parity check was still running from the first time, not sure if that's going to mess anything up. 

Server fired right up, everything seemeds normal except that both my parity drives have a UDMA CRC error count of 1 now.  I believe this is just usually a connection error, i'll keep an eye on it to see if it increases.  

 

Any ideas?  Any red flags in the diagnostics?

Diagnostic 083024.zip

Normal syslog starts overs after every boot, enable the syslog server and post that if it crashes again.

19 minutes ago, JorgeB said:

Normal syslog starts overs after every boot, enable the syslog server and post that if it crashes again.

Thats what i was worried of.  

 

I enabled the syslog server, hopefully i dont have to use it anytime soon.  

1 hour ago, JorgeB said:

Normal syslog starts overs after every boot, enable the syslog server and post that if it crashes again.

Do i need to do anything other than these settings?  

 

In the event of another crash, what information should i provide?

As mentioned in the syslog server link you need to get something recorded you either need to set up the mirror to flash option, or put your server’s IP address into the Remote Server field.

 

The mirror to flash option is the easiest to set up, but if you are worried about excessive wear on the flash drive you can put your server’s address into the Remote Server field.

49 minutes ago, itimpi said:

As mentioned in the syslog server link you need to get something recorded you either need to set up the mirror to flash option, or put your server’s IP address into the Remote Server field.

 

The mirror to flash option is the easiest to set up, but if you are worried about excessive wear on the flash drive you can put your server’s address into the Remote Server field.

i might just do both.  Is it an insane amount of writes to the flash?

 

look about right?

On 8/30/2024 at 7:03 AM, JorgeB said:

Normal syslog starts overs after every boot, enable the syslog server and post that if it crashes again.

Well, it happened again...3rd time in 3 days.  Seems like everything is fine when i go to sleep and then something bad happens overnight.  

Attached is the syslog.  

ping'ing @itimpi also

syslog-192.168.1.99.log

Edited August 31, 20241 yr by danimal86

I'm kinda grasping at straws, but could this be a hba card issue?  Looking through the log's final entry last night:
 

Quote

sd 1:0:9:0: attempting task abort!scmd(0x000000009620e53a), outstanding for 30153 ms & timeout 30000 ms

Reading through some other threads it was said that it was an issue with hba during spindown.  A parity check was going because of an un-clean shutdown, so it shouldn't have spundown any drives. My cache is dual m.2's on the mobo.  

Constant errors with multiple disks, check/replace cables for those, power and SATA.

1 hour ago, JorgeB said:

Constant errors with multiple disks, check/replace cables for those, power and SATA.

Could this be a hba issue?  I'm using a hba and a SAS expander.  

 

Any way of telling what disks are erroring?  I have a few disks

I did get UDMA CRC error counts errors on one of the unclean shutdowns but the value hasn't gone up past 1 for both drives.  

I got no other errors on the dashboard.  

 

I went ahead and ordered replacement HBA and Sas Expander cards.  I have a spare psu as well.  It would be a pain to redo every cable to the drives.  Any way to tell which ones are erroring?

Edited September 1, 20241 yr by danimal86

7 hours ago, danimal86 said:

Any way of telling what disks are erroring?

You can see in the syslog, example:

 

Aug 30 10:18:35 Tower kernel: sd 1:0:8:0: [sdl] tag#7030 CDB: opcode=0x88 88 00 00 00 00 00 67 ed a4 a8 00 00 04 00 00 00

 

sdm and sdh also have similar errors, in my experience this is not usually a HBA problem, more a cable or power issue.

 

 

5 hours ago, JorgeB said:

You can see in the syslog, example:

 

Aug 30 10:18:35 Tower kernel: sd 1:0:8:0: [sdl] tag#7030 CDB: opcode=0x88 88 00 00 00 00 00 67 ed a4 a8 00 00 04 00 00 00

 

sdm and sdh also have similar errors, in my experience this is not usually a HBA problem, more a cable or power issue.

 

 

Ok, i'm kinda stuck until the parity check completes.  Once its done i'll start messing with the cables and double check SDL/SDM/SDH drives.  What part of that error shows you that there's drive errors (just for future reference so i can look)

 

Last night, i went ahead and turned off every docker except for plex and turned off a script and any dockers that would do work overnight and the server didn't lock up over night.  Maybe i'll actually be able to finish this parity check.  

2 hours ago, danimal86 said:

What part of that error shows you that there's drive errors (just for future reference so i can look)

The same quoted line is already an error.

12 minutes ago, JorgeB said:

The same quoted line is already an error.

is it this part thats telling you there is an error:

Quote

sd 1:0:8:0:

or the entire line is what tells you that drive SDL has an error. 

just trying to figure out what to look for in the future.  

 

The complete line, and the other ones around it in the syslog.