May 1, 201115 yr As described in the subject I get this error. I have got 1 Parity disk 2T, 2* 2T, 1,5T and a newly added 1T drive. I replaced my old Files on the USB Stick as described in the upgrade description and rebooted the Server. Now all Drives beside the new one are red. What could I do? Update: I replaced the old 4.5.6 files and rebooted the server again, after this all disks are fine. Looks like that there are updates to my drivers, and thus the original drives are not known correctly anymore. What could I do to upgrade? unraidlog.txt
May 2, 201115 yr Interesting syslog! You have 5 drives on an Areca RAID controller(!), and in this 4.7 syslog, the unRAID modules are unable to read the model or serial number from any of the drives, which is why they show as missing. That is not surprising, since I don't believe Tom has ever added any support for Areca cards. If you got it working under 4.5.6, I'll bet a number of users would like to know how you did it! dgaschk: A tip for identifying Gigabyte and other motherboards - very early in the syslog, there is an ACPI table with coded entries that often point to the author of the BIOS, which sometimes helps to identify the motherboard maker. Intel boards usually indicate 'INTEL' in this area. SuperMicro boards usually have a string similar to 'C2SEA'. Gigabyte boards usually have 'GBT'. Athuruga's syslog has the string 'A M I OEM', so this is one of many boards with an AMI BIOS. Intel, SuperMicro, and Gigabyte apparently have their own BIOS developers. Below are some samples. kernel: ACPI: XSDT bfff0100 0003C (v01 A M I OEMXSDT 05000622 MSFT 00000097) AMI BIOS kernel: ACPI: FACP bfcf3040 00074 (v01 GBT GBTUACPI 42302E31 GBTU 01010101) Gigabyte motherboard kernel: ACPI: DSDT 7DD90450, 6679 (r1 C2SEA C2SEA000 0 INTL 20051117) SuperMicro motherboard kernel: ACPI: RSDT 1FE30000, 0038 (r1 INTEL D865GBF 20050307 MSFT 97) Intel motherboard kernel: ACPI: FACP 3fff30c0 00074 (v01 Nvidia AWRDACPI 42302E31 AWRD 00000000) Award BIOS, nVidia variant
May 2, 201115 yr Author Hi, This is my System: 3 HE/U 19" Industrial Server Case AIC RMC-3E2-PI-XPSS 3 x Power Supply hot plug redundant (2+1) Zippy R3G-6650P 350W 16 x HDD hot swap frame 3.5" 1 x SATA/SATA-II Backplane 1 x optical drive slim DVD/CDRW, 1 x FDD slim option (empty) 1 x Controller Areca ARC-1160ML PCI-X (for 16 HDDs) Motherboard Supermicro H8DAE-B with 2 x CPU AMD Opteron Dual Core 2.2 GHz OSA275FAA6CB + orig. AMD Heatpipe Cooler 4 GB (2 x 2 GB) RAM DDR PC3200 ECC reg. CL3 I will add the Syslog As soon As I'm at home. I run this since a halfe year now. Thx for looking into this. Regards Athu syslog-2011-05-03.txt
May 2, 201115 yr Author I just attached the syslog to my last post. Btw. Now I can see that at the end of the log the server is not able to spin the drives down, just wonder what happened now, because I never had this issue before.
May 3, 201115 yr The current syslog looks very similar to the 4.7 syslog. unRAID notes that model and serial numbers are not available for the drives but it mounts them anyway. This does not look like a driver issue but more like a change in the way unRAID identifies drives. Hopefully, someone who knows more about this will chime in. I do have some options though: 1. Select a new controller or controllers from those discussed here: http://lime-technology.com/forum/index.php?topic=12404.0. This is guaranteed to solve this problem. 2. Try to load the latest 5.0 beta. You should research how to downgrade because you'll likely need to and the procedure is not as simple when going from 5.0 to 4.x. Just make sure that you don't select FORMAT and you should be able to downgrade and get your current version working again. This option troubles me a bit because the upgrade instruction specifically require that you are operational under 4.7 before moving to 5.0. EDIT: If you do try this option, back up your current flash and start with a fresh 5.0b6a install, copy your current key to the clean install and see if the drives are recognized. If they are not, you can easily restore from the backup.
May 4, 201115 yr Author Hi, I just bought a new stick installed the 5.0 Beta and rebooted the system (I have on licence in spare, as I bought 2 at once) I did not apply the licence yet. It looks like that the hard disks are recognized, but all are unknown. I dont know how to progress... Next step I would try is 1. make a backup of my first stick 2. try to upgrade from 4.5.6 to 5.0 Beta 3. check what happens.... I hope someone else could give me also a hint. Replacing the controler makes no sence, as it would be too expensive, and I would have to replace the backplane as well. thanks for the help so far! Edit: OK after upgrading my old stick to 5.0 Beta, the normal Browser items are not visible (instead I have a broken link like: <?=$pg['Title'];?> ), but I'm able to use Port 89 and then I can see unmenu etc. In myMain I see that my 5 disks are unknown (and orange) but now there are 5 "new" disks, having a serial etc. The drives have got IDs from 7800 to 7840. If I click on the ID I could see the serial, but no manufactorer. Parts of the syslog: May 4 14:16:38 Tower emhttp: 20004d92780447810 (sdb) 1953514552 --- 1. drive May 4 14:16:38 Tower emhttp: 20004d92780447820 (sdc) 1953514552 --- 2. drive May 4 14:16:38 Tower emhttp: 20004d92780447830 (sdd) 1465138552 --- 3. drive May 4 14:16:38 Tower emhttp: 20004d92780447840 (sde) 976762552 --- 4. drive (which is empty) May 4 14:16:38 Tower emhttp: 20004d92780447800 (sda) 1953514552 --- my Parity Because of this my original drives now: May 4 14:16:38 Tower kernel: md: unRAID driver 2.1.0 installed May 4 14:16:38 Tower kernel: md: converting superblock version 0 to version 2 May 4 14:16:38 Tower kernel: mdcmd (1): import 0 0,0 May 4 14:16:38 Tower kernel: md: disk0 missing May 4 14:16:38 Tower kernel: mdcmd (2): import 1 0,0 May 4 14:16:38 Tower kernel: md: disk1 missing May 4 14:16:38 Tower kernel: mdcmd (3): import 2 0,0 May 4 14:16:38 Tower kernel: md: disk2 missing May 4 14:16:38 Tower kernel: mdcmd (4): import 3 0,0 May 4 14:16:38 Tower kernel: md: disk3 missing May 4 14:16:38 Tower kernel: mdcmd (5): import 4 0,0 May 4 14:16:38 Tower kernel: md: disk4 missing Any suggestions what I could do? Syslog_5.0_Beta.txt
May 4, 201115 yr You have a big problems as this Areca card did not "work" correctly in Unraid 4.5.6 Four of the five drives there are with different sizes. May 2 19:19:18 Tower emhttp: unRAID System Management Utility version 4.5.6 May 2 19:19:18 Tower emhttp: Copyright © 2005-2010, Lime Technology, LLC May 2 19:19:18 Tower emhttp: Pro key detected, GUID: removed May 2 19:19:18 Tower emhttp: shcmd (1): udevadm settle May 2 19:19:18 Tower emhttp: Device inventory: May 2 19:19:18 Tower emhttp: pci-0000:02:0e.0-scsi-0:0:0:0 host1 (sda) 20004d92780447800 May 2 19:19:18 Tower emhttp: pci-0000:02:0e.0-scsi-0:0:1:0 host1 (sdb) 20004d92780447810 May 2 19:19:18 Tower emhttp: pci-0000:02:0e.0-scsi-0:0:2:0 host1 (sdc) 20004d92780447820 May 2 19:19:18 Tower emhttp: pci-0000:02:0e.0-scsi-0:0:3:0 host1 (sdd) 20004d92780447830 May 2 19:19:18 Tower emhttp: pci-0000:02:0e.0-scsi-0:0:4:0 host1 (sde) 20004d92780447840 and then the sizes: May 2 19:19:19 Tower kernel: md: import disk0: [8,0] (sda) <no model> <no serial_no> offset: 32 size: 1953514568 May 2 19:19:19 Tower kernel: md1: import: scsi_inquiry (std inquiry) error: -14 May 2 19:19:19 Tower kernel: md1: import: scsi_inquiry (vpd: unit ser no) error: -14 May 2 19:19:19 Tower kernel: md: import disk1: [8,16] (sdb) <no model> <no serial_no> offset: 32 size: 1953514568 May 2 19:19:19 Tower kernel: md2: import: scsi_inquiry (std inquiry) error: -14 May 2 19:19:19 Tower kernel: md2: import: scsi_inquiry (vpd: unit ser no) error: -14 May 2 19:19:19 Tower kernel: md: import disk2: [8,32] (sdc) <no model> <no serial_no> offset: 32 size: 1953514568 May 2 19:19:19 Tower kernel: md3: import: scsi_inquiry (std inquiry) error: -14 May 2 19:19:19 Tower kernel: md3: import: scsi_inquiry (vpd: unit ser no) error: -14 May 2 19:19:19 Tower kernel: md: import disk3: [8,48] (sdd) <no model> <no serial_no> offset: 32 size: 1465138568 May 2 19:19:19 Tower kernel: md4: import: scsi_inquiry (std inquiry) error: -14 May 2 19:19:19 Tower kernel: md4: import: scsi_inquiry (vpd: unit ser no) error: -14 May 2 19:19:19 Tower kernel: md: import disk4: [8,64] (sde) <no model> <no serial_no> offset: 63 size: 976762552 The first four drives in 4.7 and 5.0b6 are recognized as: 1953514552 1953514552 1953514552 1465138552 which are the native size for 2TB and 1.5TB drives Somehow the card managed to increase the native size by 16 bytes and now for Unraid 4.7 and 5.0b these are treated as new drives and hence your "too many wrong and/or missing disks" I do not know how to proceed since you have an older motherboard without SATA slots. If you want to keep your configuration I will suggest to buy one of the supported Supermicro PCI-X cards. Stay with 4.5.6 but connect one of the four disks to the the new SM controller, then build parity, check parity, then stop down the array, connect the second drive to the SM, build parity, check parity and continue until you fix the size for all four drives. Only then you can upgrade to 4.7 or 5.0b6a and preferably keep all the drives on the SM controller (8 ports). Once you move past the 8 HDs you can reuse Areca as the card at least recognizes them at proper size (still without the serial numbers, temperature etc. - and this never worked in 4.5.6 too). Or if you do not want to buy the relatively expensive SM PCI-X 8ports card you can buy a cheap 4-ports PCI card (probably around $15) to fix the incorrect size in 4.5.6 but the downsize is once you add the third and the fourth HD to it the parity builds and checks may take a day. PS. Forgot - the parity drive should be the last one to be subjected to this procedure. And there is no proper support for this Areca Card - it appears to be usable with many limitations so proceed with caution.
Archived
This topic is now archived and is closed to further replies.