doron

When I returned home today I found that my unRAID server (5.0) has restarted (unorderly). Since the syslog is maintained in RAM, I can't see what happened. Is there anything I can do to track back what caused it? (doesn't seem to be a power issue). What is considered to be the common wisdom to address this? Link /var/log/syslog to permanent storage? rsyslog? (wouldn't have helped for sudden crash or panic anyway). Something else? Thanks for any insight.

Creating elsewhere and then moving is certainly an option. The other option is to create it on the unRAID server itself. Plus, you can choose to mount it in there and access remotely.

Okay, I have a few findings. Kinda interesting. I'll start with the bottom line. I got Truecrypt 7.1a to work nicely under unRAID 5.0. The "nokernelcrypto" option is key. The beef: a. I made the device mapper kernel module (dm-mod), which was only the beginning: to actually run, truecrypt in turn required a few other modules: dm-crypt, xts and gf128mul. Once all these were made available, there were no more kernel complaints; however, trying to mount a volume with truecrypt hung (never returned), and the mount did not complete. I could interrupt (^C), at which time I found that the work was half done - "tryecrypt -l" reports the volume as open, but it is not mounted. I didn't follow this path further, due to 2 below. b. Once you use "nokernelcrypto", and you have kernel support for the FS you're mounting, you don't actually need the device mapper (or any related kernel module for that matter). I used a truecrypt binary(!!) from the truecrypt.org, and it "just works". If your FS is NTFS, you do need the ntfs-3g or else you're stuck with r/o, but this was expected and has already been covered in this topic. c. In terms of performance, I don't know how much better would using kernel crypto have been, had I managed to make it work. I made some rough, crude measurements, and in my setup, TC adds ~25% to the wall-clock timing of copying a 800MB file. I tested with a set of zeros (dd if=/dev/zero of=/mnt/t/testfile bs=100K count=8000) and with a random file which I generated from /dev/urandom (to isolate the effect of built-in compression). Both gave similar readings. Bottom line: to mount a TC volume in unRAID, you can do this: 1. Download the "Console-Only, 32bit" installer from truecrypt.org 2. Run it (I did it on a separate system, but you can probably do it on unRAID). Select the "extract TAR" option (not "install"). untar the file. Pick up the resulting "truecrypt" binary from .../usr/bin and copy it to a permanent location (say /boot/truecrypt/ or on the array). 3. Copy that binary to /usr/bin (can do it permanently as part of "go"). 4. If your volume filesystem is NTFS, make sure ntfs-3g is installed (can do via uumenu). 5. Mount your volumes. You can either use kapperz' mount command to mount a whole drive (not part of the array), like so: /usr/bin/truecrypt -k "" --protect-hidden=no -m nokernelcrypto --filesystem=ntfs-3g -p "$TCPASS" /dev/$DRIVE /mnt/$MOUNT_POINT or mount a file container as a volume, like so: /usr/bin/truecrypt -k "" --protect-hidden=no -m nokernelcrypto --filesystem=ntfs-3g -p "$TCPASS" /mnt/disk1/$TCVOL /mnt/$MOUNT_POINT This is what I plan to do, in which case the volume is on protected storage. It is possible to create new volumes and keyfiles with the command line truecrypt, but I'm creating mine in a GUI-ed system. Hope this will help someone. Again, many thanks to kapperz for all the work and info!!

Fantastic. Thanks. I already made the dm-mod.ko before I went out to work today, will check the rest later tonight. Looking at your mount command I saw that you're mounting /dev/<dev>, which made me realize that I missed your earlier statement about the drive you're mounting not being part of the array. This is a bit different than my need (I do want this to be protected storage). I suppose I will end up mounting a file container. Ideally, I'd wish unRAID would have been able to have an encryption layer under its array rather than on top. So that you'd need to supply the password or keyfile during array mount. Hopefully, a future version feature. Thanks!

Thanks very much!! I already have a kernel build setup on a Slackware alongside unRAID. I'll try to figure out how to import this into 5.0. Just out of curiosity, you didn't try this on 5.0 because you have an alternative to TC and/or not interested anymore, or because you haven't moved to 5.0 yet? Thanks again for taking the time to document this. It is extremely helpful!

Okay, it's been like 3.5 years since the last post on this topic. A couple questions: 1. kapperz, do you have this working in unRAID 5.0? 2. Would you share details of how it was done? I'm very interested, and would prefer not to go the path of encFS or remote TC mounts. Thanks!

Got it. Thanks. I wonder why it was not included. At any rate, - has anyone compiled a CDROM kernel driver, so that I won't need to go on building a new kernel?

Hi, Does the kernel supplied with rc16c include an IDE CD driver? My server has a pretty standard, oldish, LG optical drive (E-IDE). When unRAID (rc16c) boots, it sees it during initial boot, but then it doesn't get assigned a standard device name so it's not usable (or seen by software such as VirtualBox). Is there anything I can do to resolve this? Thank for any help.

Was there any resolution to this? I'm also getting these (plus other) error messages after installing the web server plugin. Thanks!

Right. Thanks. That is what I'm experiencing. The new webgui is actually great! Then again, when I install just the SF web server (I need it for phpvirtualbox), some of the webgui breaks with assorted PHP errors. I'm presuming it's PHP 5.2 vs. 5.3 but I haven't yet delved into it. Any way to install a web server plugin which will not break the new webgui?

Thanks for your response. Following your advice, I just did just that. Redownload this file. Then installpkg (I'm trying to isolate the problem from the plugin, but I get the very same with the plugin) and then "vboxmanage extpack install...". Same result: vboxmanage hangs, never returns; "top" shows VBoxSVC eating up as much CPU as it possibly can. What do I do now?! **UPDATE** Problem resolved. Ugh. The vbox package was okay, but the extpack itself was somehow corrupt. Refreshed it, installed plugin, exrpack installed fine. Sorry for the "noise". Thanks!

Hi folks - were any of you able to install the vbox extension pack for 4.2.16? When I try to install it - be it via the plugin (thanks!!) or via a manual install of the packages - the "vboxmanage extpack install" hangs forever and does not complete. When it does, I see the process VBoxSVC eating 100% CPU and not completing. unRAID 5 rc16c. Any help would be greatly appreciated! (I also posted about it here).

Ah - you mean this? I didn't realize it was a joint venture... That'd be great. Will try it tonight. Thanks!

It seems like SimpleFeatures has some issues with the latest set of packages in rc16c. Moreover, there seems to be a clash between packages installed by unmenu and those installed by some of the plugins - case in point, OpenSSL. Any chance these can be aligned? (e.g. when OpenSSL 1.01 is installed on top of OpenSSL 0.98, things start to break). More importantly: I noticed that the revamped webGUI lacks some of the options in the stock webGUI. One example is the disk menu. When you are in the main stock menu, and click e.g. disk1, the menu contains key options re sharing via NFS and SMB. When the SF webGUI installs, those options are gone. Is any of this planned to be fixed? Thanks!!

Hi folks, I've been pulling my hair on this one for quite a while, and seem to be hitting a brick wall. I'm trying to install latest vbox on unRAID rc16c. I've tried both via the plugin and via manual install of the package from @1aina. In both cases, package install goes well but when I try to install the extension pack (either via the plugin or manually via vboxmanage), the install freezes and never completes. Checking what happens reveals that the process VBoxSVC is consuming 100% CPU and never terminates, and vboxmanage hangs with it. Killing VBoxSVC breaks the extpack install. Online search brought no solution (there were some ideas in some forums but none solved the problem in unRAID). I'm stuck here, since I must have the extpack (need RDP for the VMs). Anyone seen this? Anyone has a counter example - i.e. vbox 4.2.16 successfully installed on unRAID 5 rc16c including the extension pack? Any help would be greatly appreciated. Thanks!!

Here's a relevant piece of syslog: Aug 4 01:59:41 Tower emhttp: shcmd (371): smbpasswd -x doron Aug 4 01:59:41 Tower emhttp: _shcmd: shcmd (371): exit status: 1 Aug 4 01:59:41 Tower emhttp: shcmd (372): userdel doron Aug 4 01:59:41 Tower emhttp: _shcmd: shcmd (372): exit status: 8 Aug 4 01:59:41 Tower emhttp: shcmd (373): cp /etc/passwd /etc/shadow /etc/samba/private/smbpasswd /boot/config Aug 4 01:59:41 Tower emhttp: shcmd (374): :>/etc/samba/smb-shares.conf Aug 4 01:59:41 Tower emhttp: Restart SMB... Aug 4 01:59:41 Tower emhttp: shcmd (375): killall -HUP smbd Aug 4 01:59:41 Tower emhttp: shcmd (376): ps axc | grep -q rpc.mountd Aug 4 01:59:42 Tower emhttp: _shcmd: shcmd (376): exit status: 1 Aug 4 01:59:42 Tower emhttp: shcmd (377): rmmod md-mod |& logger Aug 4 01:59:42 Tower emhttp: shcmd (378): modprobe md-mod super=/boot/config/super.dat slots=3 |& logger Aug 4 01:59:42 Tower kernel: md: unRAID driver removed At the time of the exit status 8 from userdel (which seems to be the problem), there's a console message: "userdel: user doron is currently logged in". As far as I can tell, the user is not logged in (never logged in interactively, and no process owned by it). **Update** after a while (few hours), without an intervening reboot, problem went away (i.e. user successfully deleted). Can't explain.

Hi - using rc16c, I'm trying to delete a user I created. No matter what I try, the user is not deleted. (check the "delete" checkbox, push delete, then done - nothing). Is this a bug or am I missing something? Thanks.