Unraid slows to a crawl

[Shu]

I was monitoring a parity check run just now when I found the webGUI to be extremely slow (to the point that tabs wouldn't load even when given 5+ minutes). 

 

I looked into the log and saw this snippet repeated over and over and over again:

Aug  3 22:28:04 520unraid kernel: sd 10:0:5:0: Power-on or device reset occurred
Aug  3 22:28:05 520unraid kernel: sd 10:0:5:0: Power-on or device reset occurred
Aug  3 23:51:09 520unraid kernel: sd 10:0:6:0: attempting task abort!scmd(0x00000000dacc87bb), outstanding for 60050 ms & timeout 60000 ms
Aug  3 23:51:09 520unraid kernel: sd 10:0:6:0: [sdh] tag#2298 CDB: opcode=0x12 12 00 00 00 24 00
Aug  3 23:51:09 520unraid kernel: scsi target10:0:6: handle(0x0010), sas_address(0x4433221106000000), phy(6)
Aug  3 23:51:09 520unraid kernel: scsi target10:0:6: enclosure logical id(0x500605b0055806c0), slot(1) 
Aug  3 23:51:10 520unraid kernel: sd 10:0:6:0: task abort: SUCCESS scmd(0x00000000dacc87bb)
Aug  3 23:51:10 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  3 23:51:10 520unraid kernel: sd 10:0:6:0: Power-on or device reset occurred
Aug  3 23:51:10 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  3 23:51:10 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  3 23:51:10 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  3 23:51:10 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  3 23:51:11 520unraid kernel: sd 10:0:6:0: Power-on or device reset occurred
Aug  4 00:14:54 520unraid kernel: sd 10:0:6:0: attempting task abort!scmd(0x000000006c97dd1d), outstanding for 60132 ms & timeout 60000 ms
Aug  4 00:14:54 520unraid kernel: sd 10:0:6:0: [sdh] tag#2306 CDB: opcode=0x12 12 00 00 00 24 00
Aug  4 00:14:54 520unraid kernel: scsi target10:0:6: handle(0x0010), sas_address(0x4433221106000000), phy(6)
Aug  4 00:14:54 520unraid kernel: scsi target10:0:6: enclosure logical id(0x500605b0055806c0), slot(1) 
Aug  4 00:14:55 520unraid kernel: sd 10:0:6:0: task abort: SUCCESS scmd(0x000000006c97dd1d)
Aug  4 00:14:55 520unraid kernel: program smartctl is using a deprecated SCSI ioctl, please convert it to SG_IO
Aug  4 00:14:55 520unraid kernel: sd 10:0:6:0: Power-on or device reset occurred
Aug  4 00:14:56 520unraid kernel: sd 10:0:6:0: Power-on or device reset occurred

I attached a syslog file here (note: above, it says Aug 4th but I only found out about it today, Aug 10th. Am just realizing now that it apparently has been going on for a while...)

 

If it helps, I am using this HBA: https://www.ebay.com/itm/123916887273

syslog (3)

[JorgeB]

Please post the diagnostics.

[Shu]

5 hours ago, JorgeB said:

Please post the diagnostics.

 

520unraid-diagnostics-20220810-0123.zip

[JorgeB]

Looks more like a power/connection problem, check if the affected devices share anything, like a splitter or miniSAS cable.

 

Also a good idea to update the LSI firmware, since it's quite old.

[Shu]

33 minutes ago, JorgeB said:

Looks more like a power/connection problem, check if the affected devices share anything, like a splitter or miniSAS cable.

I do use the HBA which goes from SAS to 4 SATA. On top of that, I have this SATA power extender

 

 

33 minutes ago, JorgeB said:

Also a good idea to update the LSI firmware, since it's quite old.

Can I ask where in the diagnostic you found what firmware I'm running. Also, all I can seem to find on Broadcom's website is info for the SAS 9200-8e HBA (whereas I purchased an 8i on ebay years ago). When looking at my ebay invoice, the item's title reads "Genuine LSI 6Gbps SAS HBA LSI 9200-8i = (9211-8I) IT Mode ZFS FreeNAS unRAID US". I *did* find firmware for the 9211-8i. Do you have a better sense for which I should use? Thanks! 

Edited August 10, 2022 by Shu

[JorgeB]

16 minutes ago, Shu said:

I have this SATA power extender

These are a very bad idea.

 

16 minutes ago, Shu said:

Can I ask where in the diagnostic you found what firmware I'm running.

In the syslog, search for "fwv".

 

18 minutes ago, Shu said:

I *did* find firmware for the 9211-8i. Do you have a better sense for which I should use?

Use this one, latest release is 20.00.07.00

[Shu]

11 minutes ago, JorgeB said:

These are a very bad idea.

 

In the syslog, search for "fwv".

 

Use this one, latest release is 20.00.07.00

Darn...I'm using a Corsair 550W power supply that is not modular. Do you know of a good way to power more drives than I have ports for? I'm sure I have enough power (only have the CPU in there and an old AMD gpu (pci power only)).

[JorgeB]

You can use splitters if there's no other way, but a SATA plug should only be spitted for 2 devices, you can also use molex to SATA, those can handle more than 2 SATA disks, but if possible still all limit to 2 for splitter.

 

 

 

[Shu]

8 hours ago, JorgeB said:

Use this one, latest release is 20.00.07.00

(had an internet outage today but) I just updated the firmware! Have never done that before, exciting and nerve-wracking haha. I will monitor for the issue again

 

8 hours ago, JorgeB said:

You can use splitters if there's no other way, but a SATA plug should only be spitted for 2 devices, you can also use molex to SATA, those can handle more than 2 SATA disks, but if possible still all limit to 2 for splitter.

My PSU came with two 3-port sata cables (enough for 6 devices in total). I attached one 4-port splitter to each native sata cable (so each now supports 6) for a total of 12 across both. My PSU also has one cable that is PSU > SATA - Molex - Molex - FDD (floppy disk?). I suppose I could try the two Molex to two Molex>dual SATA splitter? That would give me 4 but still not enough... Is there a better and just completely different way people power their drives? 

[JorgeB]

7 hours ago, Shu said:

two 3-port sata cables

Each of this ports can be split in 2, so 12 total just here.