luca Posted February 11, 2011 Share Posted February 11, 2011 Yes, many people have had issues with drives missing under 5.0beta3. 5.0beta4 has that fix (I think) Did you try 5.0beta4? Yes Joe, I was trying 5beta4. Quote Link to comment
Stokkes Posted February 11, 2011 Share Posted February 11, 2011 I'm currently trying something that may fix this drive missing issue for some of us.. This isn't sanctioned by Tom, but hopefully it will fix it. A bit on the dangerous side, but I'm risking it. Quote Link to comment
RobJ Posted February 11, 2011 Share Posted February 11, 2011 I let the server run all night and most of today. This evening nothing much had changed. I forced a reboot, and am attaching the syslog. Array was stopped, and one of my 2TB drives was marked as "missing". I reverted to 4.7 and the "missing" drive is back. Any idea? I don't see anything wrong in your syslog. Are you saying that in v5.0-beta4, you could not assign sdc to Disk 6? Did you happen to capture a syslog before you rebooted? Quote Link to comment
smoldersonline Posted February 11, 2011 Share Posted February 11, 2011 I guess the "missing disk" issue (5.0-beta4) is already documented. Just to be sure, I'm attaching syslog of this morning - missing disk 7 through disk 14. syslog-2011-02-11.txt.zip Quote Link to comment
Stokkes Posted February 11, 2011 Share Posted February 11, 2011 Alright, I think I've stumbled onto a very interesting problem.. I'll do my best to explain it as I understand it. The new method by which unRAID assigns disks is now based on pci device mapping.. However, I think some hardware (motherboard related? controller card related? both?) may be changing this mapping every time the system reboots. Here's why I think this.. my disk2 is ST31500341AS_9VS0BM7F (i'm using disk2 as an example, but more disks go missing depending on reboot) From two separate syslogs.. syslog-20110210-2331: Feb 10 23:30:45 stewie emhttp: ST31500341AS_9VS0BM7F (sdb) 1465137492 => pci0000:00/0000:00:1f.2/host1/target1:0:1/1:0:1:0 . . Feb 10 23:30:45 stewie kernel: md: disk2 missing I then run "New Slots" to assign it based on the signature, which fixes the issue, then reboot again... syslog-20110210-2346: Feb 10 23:45:27 stewie emhttp: ST31500341AS_9VS0BM7F (sde) 1465137492 => pci0000:00/0000:00:1f.2/host2/target2:0:1/2:0:1:0 . . . Feb 10 23:45:27 stewie kernel: md: disk2 missing What the.....? Very interesting problem. My PCI devices are changing on reboot, which is why unRAID can't pin them down and reboots with missing disks every time. I wonder if this new mapping mechanism is inherently flawed? (Not Tom's fault though).. syslog-20110210-2331.txt.zip syslog-20110210-2346-postreboot.txt.zip disk.cfg disk-20110210-234005.cfg Quote Link to comment
luca Posted February 11, 2011 Share Posted February 11, 2011 I let the server run all night and most of today. This evening nothing much had changed. I forced a reboot, and am attaching the syslog. Array was stopped, and one of my 2TB drives was marked as "missing". I reverted to 4.7 and the "missing" drive is back. Any idea? I don't see anything wrong in your syslog. Are you saying that in v5.0-beta4, you could not assign sdc to Disk 6? Did you happen to capture a syslog before you rebooted? Like I said, all seemed fine after booting into 5beta4, I was able to assign all the drives, and start the array. The problem started upon changing the file permissions (step 4 of Tom's release notes). After that the connection via browser became unresponsive, and also the shares were not accessible anymore (I could still access the disks). After many hours unraid was still unresponsive, so I simply rebooted. The syslog that was posted is after the server was rebooted and I had access to the web console again. I did not try to (re)assign disk6, sorry. I was in a hurry to get the server running and I just reverted to 4.7. Could I have captured the syslog before rebooting? Please explain how. I can try this again tomorrow, to see if I can reproduce this problem. Quote Link to comment
johnodon Posted February 11, 2011 Share Posted February 11, 2011 So, it appears that the webgui format is off for 1920 x 1080 rez's. My desktop machine is running 1920 x 1080, but I don't get a scroll bar at the bottom, and everything lines up perfectly. I do get a scroll bar on the right, even though it's unnecessary and has no slider. Here are 2 screeshots: 1680 x 1050 (which displays correctly with lines centered) 1920 x 1080 (which has the lines shifted and scroll bar at bottom) Quote Link to comment
Joe L. Posted February 11, 2011 Share Posted February 11, 2011 What specific browser are you using? I'm guessing it has more to do with that than anything else. I have the same horizontal scrollbar if I use IE 7 and expand to full screen mode on my 1920x,1200 monitor, but the horizontal scrollbar is not there when using Firefox or Google-Chrome. (and the screen looks perfect) It is probably an error with IE not following the CSS standards but one Microsoft made up instead. Quote Link to comment
johnodon Posted February 11, 2011 Share Posted February 11, 2011 What specific browser are you using? I'm guessing it has more to do with that than anything else. I have the same horizontal scrollbar if I use IE 7 and expand to full screen mode on my 1920x,1200 monitor, but the horizontal scrollbar is not there when using Firefox or Google-Chrome. (and the screen looks perfect) It is probably an error with IE not following the CSS standards but one Microsoft made up instead. In every case I am using IE8. John Quote Link to comment
PeterB Posted February 11, 2011 Share Posted February 11, 2011 It is probably an error with IE not following the CSS standards but one Microsoft made up instead. That is one of the problems. Microsoft don't follow standards - they want to set their own standards and, ultimately, everyone else has to follow! Quote Link to comment
larson Posted February 11, 2011 Share Posted February 11, 2011 Alright, I think I've stumbled onto a very interesting problem.. I'll do my best to explain it as I understand it. The new method by which unRAID assigns disks is now based on pci device mapping.. However, I think some hardware (motherboard related? controller card related? both?) may be changing this mapping every time the system reboots. Here's why I think this.. my disk2 is ST31500341AS_9VS0BM7F (i'm using disk2 as an example, but more disks go missing depending on reboot) From two separate syslogs.. syslog-20110210-2331: Feb 10 23:30:45 stewie emhttp: ST31500341AS_9VS0BM7F (sdb) 1465137492 => pci0000:00/0000:00:1f.2/host1/target1:0:1/1:0:1:0 . . Feb 10 23:30:45 stewie kernel: md: disk2 missing I then run "New Slots" to assign it based on the signature, which fixes the issue, then reboot again... syslog-20110210-2346: Feb 10 23:45:27 stewie emhttp: ST31500341AS_9VS0BM7F (sde) 1465137492 => pci0000:00/0000:00:1f.2/host2/target2:0:1/2:0:1:0 . . . Feb 10 23:45:27 stewie kernel: md: disk2 missing What the.....? Very interesting problem. My PCI devices are changing on reboot, which is why unRAID can't pin them down and reboots with missing disks every time. I wonder if this new mapping mechanism is inherently flawed? (Not Tom's fault though).. I am having missing disk issues also. See my post (with syslogs) http://lime-technology.com/forum/index.php?topic=10896.0 I have to reassign my array after reboot. Don't know if you mean PCI as in the old bus or as in a way of working internally in Linux, but my devices are all on either on-board chipset or PCI-E Adaptec 1430. And I am running beta4. Quote Link to comment
bcbgboy13 Posted February 11, 2011 Share Posted February 11, 2011 Question for Stokkes: What motherboard are you using - model, hardware revision if any and BIOS. Quote Link to comment
Stokkes Posted February 11, 2011 Share Posted February 11, 2011 Question for Stokkes: What motherboard are you using - model, hardware revision if any and BIOS. Asus P5K-E.. not sure of the bios revision. Quote Link to comment
Stokkes Posted February 11, 2011 Share Posted February 11, 2011 I am having missing disk issues also. See my post (with syslogs) http://lime-technology.com/forum/index.php?topic=10896.0 I have to reassign my array after reboot. Don't know if you mean PCI as in the old bus or as in a way of working internally in Linux, but my devices are all on either on-board chipset or PCI-E Adaptec 1430. And I am running beta4. Can you upload (in that thread) your disk.cfg and maybe a screen shot of the main unRAID menu after a fresh reboot of your system (with the missing disks indicated)? Quote Link to comment
hellbringer Posted February 12, 2011 Share Posted February 12, 2011 Since the download is quite slow (50kb/s here), I created a torrent file for this, I hope that's allright? See link, it's using Open trackers (linuxtracker.org, openbittorrent.com & publicbt.com) and also the weburl from the site. Still working to get it on Linuxtracker.org, somehow theyr upload function is broken. http://www.2shared.com/file/Hic-H_n4/unRAID_Server_50-beta4_AiOzip.html Quote Link to comment
RobJ Posted February 12, 2011 Share Posted February 12, 2011 Since the download is quite slow (50kb/s here), I created a torrent file for this, I hope that's allright? See link, it's using Open trackers (linuxtracker.org, openbittorrent.com & publicbt.com) and also the weburl from the site. Still working to get it on Linuxtracker.org, somehow theyr upload function is broken. http://www.2shared.com/file/Hic-H_n4/unRAID_Server_50-beta4_AiOzip.html For safety whenever downloading the unRAID distribution, through the links above or any other source, torrent or not, please make sure that the MD5 or SHA1 of the downloaded zip file matches the following: MD5: 484E006E17B7A268A9B49799AE3F4507 SHA1: A360BBF50BB258D23E11B8CDEEC796B0443563D3 Free tools for Windows: MD5check (below), HashOnClick (MD5 & SHA1) by 2BrightSparks (maker of SyncBack) Quote Link to comment
hellbringer Posted February 12, 2011 Share Posted February 12, 2011 Since the download is quite slow (50kb/s here), I created a torrent file for this, I hope that's allright? See link, it's using Open trackers (linuxtracker.org, openbittorrent.com & publicbt.com) and also the weburl from the site. Still working to get it on Linuxtracker.org, somehow theyr upload function is broken. http://www.2shared.com/file/Hic-H_n4/unRAID_Server_50-beta4_AiOzip.html For safety whenever downloading the unRAID distribution, through the links above or any other source, torrent or not, please make sure that the MD5 of the downloaded zip file matches the following: 484E006E17B7A268A9B49799AE3F4507 Good point. You can check the MD5 with the windows tool MD5check, found here: http://www.softpedia.com/get/System/File-Management/MD5-Check.shtml Quote Link to comment
GK20 Posted February 12, 2011 Share Posted February 12, 2011 For safety whenever downloading the unRAID distribution, through the links above or any other source, torrent or not, please make sure that the MD5 of the downloaded zip file matches the following: 484E006E17B7A268A9B49799AE3F4507 Good point. You can check the MD5 with the windows tool MD5check, found here: http://www.softpedia.com/get/System/File-Management/MD5-Check.shtml MD5 is now known to be "weak" for anti-tampering in message digest, most are moving to SHA1 or SHA2, as this article indicated. http://tools.cisco.com/security/center/viewAlert.x?alertId=17341 Many open source packages distributing online now use either SHA1 or PGP signature. Quote Link to comment
RobJ Posted February 12, 2011 Share Posted February 12, 2011 Alright, I think I've stumbled onto a very interesting problem.. I'll do my best to explain it as I understand it. The new method by which unRAID assigns disks is now based on pci device mapping.. However, I think some hardware (motherboard related? controller card related? both?) may be changing this mapping every time the system reboots. Here's why I think this.. my disk2 is ST31500341AS_9VS0BM7F (i'm using disk2 as an example, but more disks go missing depending on reboot) From two separate syslogs.. syslog-20110210-2331: Feb 10 23:30:45 stewie emhttp: ST31500341AS_9VS0BM7F (sdb) 1465137492 => pci0000:00/0000:00:1f.2/host1/target1:0:1/1:0:1:0 . . Feb 10 23:30:45 stewie kernel: md: disk2 missing I then run "New Slots" to assign it based on the signature, which fixes the issue, then reboot again... syslog-20110210-2346: Feb 10 23:45:27 stewie emhttp: ST31500341AS_9VS0BM7F (sde) 1465137492 => pci0000:00/0000:00:1f.2/host2/target2:0:1/2:0:1:0 . . . Feb 10 23:45:27 stewie kernel: md: disk2 missing What the.....? Very interesting problem. My PCI devices are changing on reboot, which is why unRAID can't pin them down and reboots with missing disks every time. I wonder if this new mapping mechanism is inherently flawed? (Not Tom's fault though).. I am having missing disk issues also. See my post (with syslogs) http://lime-technology.com/forum/index.php?topic=10896.0 I have to reassign my array after reboot. Don't know if you mean PCI as in the old bus or as in a way of working internally in Linux, but my devices are all on either on-board chipset or PCI-E Adaptec 1430. And I am running beta4. I have always hoped that someday Tom would change his initial disk assignment process to ignore any hardware ID's and just link drive serial numbers to disk numbers. It would remove one of the last hardware dependencies, and make motherboard changes and major kernel upgrades much easier, no longer requiring the user to record the disk assignments and reassign them, after the hardware or software change. Disk.cfg would change from something like this: # Disk configuration spindownDelay=30 queueDepth=1 spinupGroups=yes md_num_stripes_default=1280 md_write_limit_default=768 md_sync_window_default=288 parity=pci-0000:03:00.0-scsi-0:0:0:0 disk1=pci-0000:00:05.1-scsi-0:0:0:0 disk2=pci-0000:00:05.2-scsi-0:0:0:0 To something like this: # Disk configuration spindownDelay=30 queueDepth=1 spinupGroups=yes md_num_stripes_default=1280 md_write_limit_default=768 md_sync_window_default=288 parity=JP2921HQ1Z8HVA disk1=5QG0D11W disk2=S0MUJ1KP206409 Except for network settings and addons that require network configuration, an unRAID server becomes very portable. I imagine you could grab a 5 disk drive enclosure with port multiplier support and your flash disk, and while traveling around the country, could attach it to any eSATA port (with port multiplier support) and boot from your flash, and after adjusting the network settings, have your entire 8 terabyte unRAID server available. Of more interest to most unRAID users is the fact that hardware upgrades and version upgrades are more trouble-free, with less support calls to Limetech and the forum gurus. Quote Link to comment
limetech Posted February 12, 2011 Author Share Posted February 12, 2011 I have always hoped that someday Tom would change his initial disk assignment process to ignore any hardware ID's and just link drive serial numbers to disk numbers. It would remove one of the last hardware dependencies, and make motherboard changes and major kernel upgrades much easier, no longer requiring the user to record the disk assignments and reassign them, after the hardware or software change. Disk.cfg would change from something like this: # Disk configuration spindownDelay=30 queueDepth=1 spinupGroups=yes md_num_stripes_default=1280 md_write_limit_default=768 md_sync_window_default=288 parity=pci-0000:03:00.0-scsi-0:0:0:0 disk1=pci-0000:00:05.1-scsi-0:0:0:0 disk2=pci-0000:00:05.2-scsi-0:0:0:0 To something like this: # Disk configuration spindownDelay=30 queueDepth=1 spinupGroups=yes md_num_stripes_default=1280 md_write_limit_default=768 md_sync_window_default=288 parity=JP2921HQ1Z8HVA disk1=5QG0D11W disk2=S0MUJ1KP206409 Except for network settings and addons that require network configuration, an unRAID server becomes very portable. I imagine you could grab a 5 disk drive enclosure with port multiplier support and your flash disk, and while traveling around the country, could attach it to any eSATA port (with port multiplier support) and boot from your flash, and after adjusting the network settings, have your entire 8 terabyte unRAID server available. Of more interest to most unRAID users is the fact that hardware upgrades and version upgrades are more trouble-free, with less support calls to Limetech and the forum gurus. You got your wish in 5.0-beta5 Quote Link to comment
peter_sm Posted February 12, 2011 Share Posted February 12, 2011 Is it easy to go back to 4.7? I thinking about the file permission on my share that is set from V5. Need to have a better NFS performance. //Peter Quote Link to comment
BRiT Posted February 12, 2011 Share Posted February 12, 2011 The file permissions set from the 5.0 beta newperms scripts is supposed to work perfectly fine in unRAID 4.7. You need to overwrite your flash drives bzroot and bzimage files and restore your backup copies of your config folder to fully revert to unRAID 4.7. If you do not have backup copies, of your config folder, that you should have made before upgrading to 5.0 betas, then you may need to reassign the drives in your array since the super.dat file is not backwards compatible and some .cfg files may have their internal data used differently. Quote Link to comment
peter_sm Posted February 12, 2011 Share Posted February 12, 2011 Thanks, I have my USB completed Back upped from 4.7 Quote Link to comment
limetech Posted February 12, 2011 Author Share Posted February 12, 2011 Is it easy to go back to 4.7? I thinking about the file permission on my share that is set from V5. Need to have a better NFS performance. //Peter Are you saying read/write performance under NFS is different in 5.0 vs. 4.7? Quote Link to comment
JackBauer Posted February 12, 2011 Share Posted February 12, 2011 You got your wish in 5.0-beta5 Does this mean that I could pull out all my HD's, then randomly re-insert them into my 5in3's... in a totally random order... And the system would work? Because that would be awesome. (To never have to worry about which drive goes where) Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.