History of ASR-5405Z Configuration Notes
| Version 2 | Current version | |
|---|---|---|
Initial dudk.ai/Llama4 discussion ((Duck.ai Llama4 - Hardware detection on Linux server - 20260421)) | I am starting to get to understand just where 'AI' can help although you need the right chat bot to be successful OK The starting point here is that I'd picked up an 8Tb SAS disk drive as the price was good not spotting that it was not SATA. As it turns out it has probably done me a favour as fixing the problem has provided another 4 disk connections on the media server, and since there is still an empty slot in the hot swap cage I can now fill up the machine. When the disk first arrived I simply plugged it in and it powered up fine, but I could not see it, at least not with the tools I was initially using. It was not until I tried to plug it into the USB SATA box to try that path that I found the connectors are actually different. Then I picked up on the mistake. A couple of suggestions were made on a suitable extra controller, but all were long lead time from China, and I get impatient ;) but digging around, the ASR-5405Z popped up in the refurbished area on eBay at only £14 delivered. At that price it was a little of a no-brainer, and a quick check confirmed that it should work on Tumbleweed. Which actually it does. The first problem was while some of the SCSI management tools were showing that there was an extra disk, the card was obviously in a mode where it was still trying to establish it's setup, and the message I was getting from rescan-scsi-bus.sh was actually flagging that. What I should have concentrated on initially was the management tool arcconf which is not available from openSUSE repos, so Claude pointing me to Microchip was a useful poke. arcconf_v3_07_23980.zip includes a .rpm package which loaded without a problem ... well after I overrid the earlier arcconf_B26540.zip install of version 4. The discussion with Claude I think confirms that V4 version was hanging because it was awaiting for the controller to be ready, while V3 simply said initially that the controller did not exist. A later retry obtained the good news that there was a controller to play with, and the arcconf tools are essential to connect the drives to the controller and create the links to /dev/sdX. Confirming the controller is seendmesg | grep aacraid lspci | grep -i adaptec Installing arcconfDownload v3 from Microchip — v4 hangs waiting for controller, v3 responds immediately: arcconf_v3_07_23980.zip from microchip.com rpm -i arcconf*.rpm Check controller and disk statusarcconf getconfig 1 PD Look for disk showing as Ready before proceeding. Create JBOD to expose disk as block devicearcconf create 1 jbod 0 0 noprompt Where `0 0` is channel and device ID from getconfig output. Verify block device createdlsblk dmesg | tail -20 sg_map -x Partition and formatparted /dev/sdX mklabel gpt parted /dev/sdX mkpart primary xfs 0% 100% mkfs.xfs /dev/sdX1 Mount and add to fstabmkdir -p /media3 mount /dev/sdX1 /media3 blkid /dev/sdX1 # Add UUID to /etc/fstab Adding subsequent disksSame process — wait for Ready status in `arcconf getconfig 1 PD` before creating JBOD! 😄 Extra NotesXFS vs ext4 note — worth a line explaining why XFS for large media files and the 400GB saving on 8TB from no reserved blocks. The PQual=1 explanation — one line noting that "LU temporarily unavailable" is normal while the controller initialises, just wait it out rather than panicking. Developed with Claude AI assistance - Anthropic - April 2026 |