hanabi

Simple question: does anyone have a CRC32 executable that will run under unRaid? I know MD5 is available but I'd prefer CRC32 for a couple of reasons: (1) having 25 TBytes of data to periodically check, I'd like a faster hash (I can live with a 32-bit hash). (2) I currently check my unRaid files using the Window's software Exactfile that can calculate CRC32. But this is calculated across the LAN (slower, and reduces LAN bandwidth available for other needs) so an executable that runs under unRaid would be ideal. Does anyone have such a executable that I can beg, borrow, or steal? It would need to be able to gracefully handle large files (e.g. 50 GByte ISOs). Most of the C implementations on the net I've found require the entire file data placed in a memory buffer, so only practical for smallish files. Thanks in advance.

Thanks for the replies guys. So, in summary: 1. stop the array 2. select Utils->New Config 3. unassign the parity drive 4. assign the (unassigned) parity drive to a data drive slot 5. start the array 6. wait for formatting to complete. If there are no more comments I'll give this a try tomorrow. <<However ... you DO realize that removing the parity protection from your array means you will no longer be fault-tolerant >> Yes, I understand that. I now have a second NAS with a complete backup of my data. Thanks again Hanabi.

Using unRAID Server Pro version: 5.0-rc5. My current unRaid array has 25TB 10 disks plus a 4TB parity drive. I recently built another 25TB NAS using Drive Bender and copied the entire unRaid array to it, so I now have a redundant system. I'm thinking there is now no need to keep the parity drive and its associated slow write speed, and would like to use it as a data drive in unRaid. Could someone please offer suitable advice as to the steps required to make a parity drive a data drive in unRaid? Here's what I think I should do: 1. stop the array 2. unassign the parity drive 3. assign the parity drive to a data drive slot 4. ? ? ? automatic zeroing/formatting by unRaid of the old parity drive to make it a data drive? 5. start the array Is it preferable to run the pre-clear script on the parity drive prior to step 3? It was pre-cleared when initially inserted into the array. A second, unrelated question if I may. Is there an unRaid utility/add-on that will list the files that are possibly corrupt when a parity error is detected? My understanding is that, when a parity error is detected, unRaid cannot determine which file/disk has the error, so it simply corrects the parity drive. I have run CRC/MD5 checks of all my files, so when a parity error is detected it would be nice if I could check for corruption just those files that may be corrupt, rather than checking every file on the array. When it comes to unRaid I am still quite a beginner so I appreciate any help with the above questions. Hanabi

Thanks for your reply dgaschk. This is a little nerve-wracking. I am 99% sure but I want to be certain. Is there a way to confirm which disk is parity? Apparently the parity drive does not have a file system so I was thinking of individually mounting each drive using the mount command from a telnet window and confirming the data disks that way. Not sure of the correct command though. Perhaps: mkdir "/tmp/mountpoint" mount -r -t reiserfs /dev/sdX1 "/tmp/mountpoint" and then determine which drive cannot be mounted or has no data. Does that sound ok? So if I do that there will be no corruption of the data drives? I assign all drives as data and then click "start", correct? Leave "maintenance mode" unchecked? Then after finding the unformatted drive, stop the array, assign unformatted drive to parity, then start the array? Sorry to be a NOOB. Thanks again for your guidance.

Ok I might have screwed up badly here. Running unRaid 5.0-beta12. One of my seven data disks suddenly died (WD 3TB EARS), so I searched for the procedure to remove a drive and rebuild parity. I came across this link http://lime-technology.com/wiki/index.php/FAQ_remove_drive which looks official enough. Followed the procedure: Stop the array; login via telnet and type initconfig. So far so good. The procedure states "the process of parity calculation on the remaining assigned and working drives will begin when you next start the array". Ok, so I powerdown, remove the faulty drive, and boot (was this my mistake/misunderstanding??), expecting that the remaining drives would be still assigned and perhaps parity being calculated. But no, all drives are now unassigned. And silly me, I forgot to take a snapshot of the disk assignment screen. I think I know which disk was parity, but I'm not sure of the order of the data disks. So, am I sunk, or is there a way to recover my data (around 18 TB)? Help, someone, please?

I also have the Sapphire E350M1. Fantastic board. No Realtek NIC so no LAN problems. Slow transfers of about 15MB/s when using unRaid cp or rsync to transfer files from my Windows box to unRaid. But 45MB/s using Windows box to do the same transfers. This is with no parity installed. Parity check on an old SATA drive was at 75MB/s. The only problem with this motherboard is that the first two 2 of 6 SATA ports do not recognise my Hitachi 3TB hard drives, although they recognise smaller drives. May just need a new bios from Sapphire. Three other ports work fine with 3TB drives. Haven't checked the 6th port yet. Overall I'm more than pleased so far. Using unRaid 5.0b10. Just an update. The problem with hard disk detection was caused by a bad pin on a power cable. I now have a total of 6 3TB Hitachi drives working in unRaid. Used an e-SATA to SATA cable to mount the 6th drive. That gives me 15TB + parity without using any SATA/SAS cards! I can't recommend this motherboard enough

Hi SilverRubicon, Regarding your Asus E35M1 motherboard: "I have the pro version of this board and it's been humming right along for months. Running 5.0b10 at the moment. Not sure why it's so hit or miss for others." Are you using the on-board NIC (with no problems)? Do you have any cards in the PCI-e slots? I'm also considering this board but worried about the problems with the on-board NIC and interrupts. Wondering whether the interrupt problems only occur when adding PCI-e cards, i.e., if running just a stock board with no ad-on cards perhaps the NIC and interrupt problems will disappear(?) By the way, there is also a Sapphire E350 board too, with 6 SATA posts, and it doesn't have a Reaktek NIC! I have one but it is having problems recognising 3TB (Hitachi) hard disks on ports 0 and 1. Smaller disks are ok though, and ports 2 to 4 work fine with 3TB drives. Haven't tried port 5 yet. No NIC problems at all. Overall a fantastic motherboard for low-power unRaid users. Just an update. The problem with hard disk detection was caused by a bad pin on a power cable. I now have a total of 6 3TB Hitachi drives working in unRaid. Used an e-SATA to SATA cable to mount the 6th drive. That gives me 15TB + parity without using any SATA/SAS cards! I can't recommend this motherboard enough Cheers.

As per the subject title. I have a LaCie 2BigQuadra external eSATA-2 hard drive (contains 2x2TByte disks) partitioned as 2TB RAID 0 and 1TB RAID 1. Is it possible to mount such partitions under unRaid 5.0b10? I've tried mouning both sdd1 and sdd2 using the command line stated in http://lime-technology.com/forum/index.php?topic=10408.0, basically: mount -r -t hfsplus /dev/sdd1 /mnt/imported_files but this failed probably due to the wrong partition type I guess (hfsplus): root@Tower:~# mount -r -t hfsplus /dev/sdd1 /mnt/tmpdrive1 mount: wrong fs type, bad option, bad superblock on /dev/sdd1, missing codepage or helper program, or other error In some cases useful info is found in syslog - try dmesg | tail or so unRaid detects the drive and partitons: Aug 12 10:27:05 Tower kernel: ata6.00: ATA-6: eSATA-2 LaCie 2BigQuadra, 0, max UDMA/133 Aug 12 10:27:05 Tower kernel: ata6.00: 5860392464 sectors, multi 0: LBA48 Aug 12 10:27:05 Tower kernel: ata6.00: configured for UDMA/133 Aug 12 10:27:05 Tower kernel: sd 6:0:0:0: [sdd] 5860392464 512-byte logical blocks: (3.00 TB/2.72 TiB) Aug 12 10:27:05 Tower kernel: sd 6:0:0:0: [sdd] Write Protect is off Aug 12 10:27:05 Tower kernel: sd 6:0:0:0: [sdd] Mode Sense: 00 3a 00 00 Aug 12 10:27:05 Tower kernel: sd 6:0:0:0: [sdd] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA Aug 12 10:27:05 Tower kernel: sdd: sdd1 sdd2 Aug 12 10:27:05 Tower kernel: sd 6:0:0:0: [sdd] Attached SCSI disk Aug 12 10:27:05 Tower emhttp: eSATA-2_LaCie_2BigQuadra_JK11C1YAJEARBV (sdd) 2930196232 "cat /proc/partitions" returns: 8 48 2930196232 sdd 8 49 1953462272 sdd1 8 50 976731908 sdd2 "fdisk -l" returns: WARNING: GPT (GUID Partition Table) detected on '/dev/sdd'! The util fdisk doesn 't support GPT. Use GNU Parted. Disk /dev/sdd: 3000.5 GB, 3000520941568 bytes 255 heads, 63 sectors/track, 364792 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0xe81254e7 Device Boot Start End Blocks Id System /dev/sdd1 1 267350 2147483647 ee GPT Assuming this is even possible can someone help me out with the correct mount command? Thanks in advance.

The manual's use of the word "intended" is what prompted my email to tech support for clarification. Their reply, "SATA card will not work on it" seems very clear, but given other MBs have "video-only" slots yet work with non-graphics cards I remain optimistic Another unRaid member, MTA99, has this Sapphire MB. I've emailed and I'm waiting for his/her reply. If I go ahead with the purchase I'll certainly post the results. I'd still like to hear the experiences of other members whose MBs have "video-only" PCI-e slots.

After a quite frustrating search for a new motherboard as a base for a new unRaid system I finally found one that I thought might fit the specs I prefer: Low power MB with preferably built-in CPU and video; Non-RealTek NIC USB3 One PCI-e slot (at least x4 speed) to take a AOC-SAS2LP-MV8 eSATA (or a 2nd PCI-e slot to take an eSATA card) The Sapphire Pure Fusion Mini E350 http://www.sapphiretech.com/presentation/product/?cid=2&gid=1047&sgid=1045&pid=1034&psn=&lid=1&leg=0 meets the above spec, or so I thought. Pre-purchase, I downloaded the manual and notice that it states [pre]Note: This PCI-Express x16 slot only supports x4 bandwidth and is intended only for use by a graphics card.[/pre] What the??? The MB already has built in video, yet they limit the only full-size PCI-e slot to video only? That doesn't sound right. Well I emailed Sapphire tech support (replied in under 24 hours; good job!) and the reply read: [pre]Yes this PCIE slot is only intended for graphics card. No SATA card will not work on it. [/pre] Well, not wanting to give up too soon I searched for threads discussing this very issue (using other cards in "video only" slots) and came up with this: http://www.tomshardware.com/forum/286165-30-slot-video which suggests in general other cards can be used; not just video cards. So my question is, what is the experience of other system builders who have MBs with such "video-only" PCI-e slots? Is it generally the case that they can be used with non-video cards too? The MB is cheap enough to buy and test but if the general consensus is "pass" then I guess I'll have to search for another MB (any advice on which one??). Cheers.

Did you use the PCI-e 16x slot for anything, and in particular, an SATA card? I also want to use this MB for unRaid but the manual and Sapphire tech support states the x16 PCI-e slot can only be used with a video card. I want to use it with a Supermicro AOC-SASLP-MV8 if possible. Thanks in advance.

Rajahal, will the Supermicro X7SLA-H-O recognise the full capacity of the new 3TB hard drives? I'm also looking for a new low power MB, for use with 3GB drives.

Thanks. I ended up having to delete secrets.tdb in /etc/samba/private/ and then rebooting to get Samba up and running again. Ross

Thanks Joe. That problem is solved. I ran checkdisk and let it fix the flash file system (one file corrupt). I then deleted super.dat, booted, assigned my parity disk (again) and hit the start button. Parity check commenced, super.dat was then re-written by the system, and parity check later completed with zero errors. I then rebooted. Super.dat looks intact this time and parity assignment wasn't dropped this time, so things look ok. Now one other problem has reared its ugly head: no shares appearing on the network. Tower emhttp: shcmd (35): killall -HUP smbd (Minor Issues) Tower emhttp: _shcmd: shcmd (35): exit status: 1 (Other emhttp) I'll start a new thread about this problem if I can't fix it. Ross

A little further info. I rebooted to see if the super.dat file could be read by the OS. No go; I still see the "could not read superblock" error in syslog. An ls -l of the /boot/config directory shows super.dat to be a 0 byte file, last updated yesterday (about 19 hours ago), not today when I rebooted after the parity sync finished. I'm not sure if the timestamp corresponds to when I started the parity sync (I rebooted immediately prior to the parity sync). There is a super.old file of 4096 bytes with an old timestamp from last month. Doing a search of the forum I see another thread describing a similar situation to mine but with a different solution: http://lime-technology.com/forum/index.php?topic=9438.0 No need for any initconfig command. Just delete the corrupt super.dat, boot, and hit the start button. Is either solution ok? Ross