Ford Prefect

Anyone observed CPU and SYS temps during usage? The mobo appears to be a fanless design (I do not think a fan will fit on the CPU inside the 1U case and it is a supermicro standard case for this board). During some test over the last days, I found that CPU temps won't exceed 38 degrees celsius, while SYS temps are up to 52 degrees easily....there is not much room (maybe 2-3 degrees) left until the warning level. CPU started at 15 deg. while SYS started at 36 deg after powerup. This is shot after 45min uptime, see here what I mean: CPU started at 15 deg. while SYS started at 36 deg after powerup. This is shot after 45min uptime The northbridge gets quite hot, but even when mounting a fan on it, the SYS temps are staying on hot level. The whole case gets quite warm from underneath, right where the mobo is mounted. ..all sensor readings coming from the IPMI sensor interface, so I think they are accurate. I know it all depends on ambient temps and such...what makes me nervous is the big difference in the temps for CPU and SYS and that headroom level for both is not equal. I am actually worried to have it run 24x7 that way. Here's my setup again: edit: inserted screenshot of IPMI sensor view

Welcome and keep us posted on how you get along. The board might be picky with RAM, especially with 800MHz S0-DIMMs. Do not forget to jumper your EARS drives (J7-J8) before you do anything else with them!

Thanks George! ...they now ship internationally and the price has been reduced a good deal as well (from around CAD120 to roughly a 100)...I am going to order NOW

Ha!...OK, that's a usecase and feature I can understand

Are you running their kernels and VE from their testing branch? I am not sure where they stick to plain lenny and where they put in extra things. My PVE host runs kernel "2.6.32-2-pve" from testing while one of my stock lenny VM runs "2.6.26-2-486 on i686" Yes, I know of that problem...had the same with my VM running Openfiler...the OF-Gui would simply not parse the list of devices and partitions other than ide and scsi. Under OF a simple change to a script would do the trick...don't know how emhttp is accessing the info...isn't it based on PHP somewhere down the line? Yes, I have done the same with my OF installation...works OK. Yes, you're right...but trends for home use go towards more green (Atom-based) solutions, I'd think....so here you do not have a single chance to do any of that with them. Some AMD Opterons support that Feature (special version of their IOMMU) I think, but I have never seen one. Keep us posted on how you play along. I am intending to swap my setup from AMD64 based proxmox host and Openfiler NAS(VM) into an Atom based native unRaid host and run my additional VMs under virtualbox on it.

Don't know about setting the serial...what distro are you running your kvm host on and which kernel and kvm/qemu version are you using? I found that some features depend heavily on the right combination of the three. Did you try and build a custom kernel with virtio-net and -block drivers for the unRAID guest? Performance should increase a lot compared to the standard emulation...

hmm, might be a silly question to ask... ...but are you happen to have a shell/terminal session active to your unRAID box, with i'ts PWD pointing to one of the disks/shares?

...the Kingston HyperX version is also fine. I am running a 2x2GB Kit of Kingston HyperX SO-DIMM Kit 4GB PC2-5300S CL4-4-4-12 (DDR2-667) (KHX5300S2LLK2/4G) ...cannot tell if the setup would benefit from the CL4 memory compared to the standard CL5. I bought them, because they were actually cheaper than the standard version at that time.

IPMI is awsome With the IPMI version of the board, you do not need a CDROM drive. You can simply mount a local CD-Drive or ISO (local means from the client where you run the IMPI Viewer/Tool) to the box...it will appear as a USB CDROM drive. This works with their IMPI Viewer/Tool under Win and Linux. ...I will never, ever go without that feature again for a future build

Hmm...I am not sure if I understand that right... I would agree on the feature of "knowing each other" in terms of name resolution for easy connectivity as reasonable. In terms of "configuration", which I would understand as a feature to provision each unRaid install locally with the right config, I would think that this is unreasonable. The first can be achieved by setting up your network DNS and IP provisioning and would not require to change the config of a single unRAID instance if changes occur (when using DHCP). The second would call for central maintenance, deployment and provisioning infrastructure (what happens in a scenario where TOWER3 should be added later...this will result in a need for change of TOWER1 and TOWER2, wouldn't it?)

Yes, that's exactly the type of riser I meant. But I learned that riser cards could get tricky because most boards tend to be picky about it. So a riser from a different source as the board manufacturer is always a risk and a ribbon based riser is subject to environmental EMC conditions, so risks for side effects are even higher. ...hopefully the X7SPE-H(F) addresses this problem. The easiest way to tell would be to check if Supermicro would offer a riser as an option with the barebone (which they do not do for barebones based on X7SPA).

...but what you used is a feature. Putting a hostname into parts of the system that would only allow domain-names is just satisfying a system the wrong way. So the behaviour you describe is only to be expectd Put hostnames, which should resolve locally into /etc/hosts and you are fine...

I don't know that symlinks are supported via smb/cifs. I'm sure symlinks are supported via NFS and I'm sure that it would work with disk shares. OK, than my strategy would be to migrate locally - possibly using mc - on unRAID to a disk share and fix/rollut nfs mounts for my linux-clients instead of smb/cifs.

Thanks for the hint Joe L. My Source-Server seems to be fine (it is an up2date openfiler NAS). The DIR in question has been created with a linux client, mounting the OF-NAS via cifs, by running a slackware mirror script with the mountpoint being the target of that script.

You have a good point there, starcat. The linux/gnome box in question is a laptop and I am connected via WiFi, which in turn is controlled by the NetworkManager. The reason is, upon shutdown, gnome/X11 gets shutdown first and with it the NetworkManager and with it the WiFi link...letting nfs mounts hang in the air, which in turn prevents the box from a proper and fast shutdown. With nfs I tried to change boot- and stop-orders and automounts but to no permanent success. smb/cifs mounts are accessible from the Gnome panel and they seem to be unmounted prior to shutdown as well. ...at least they are not giving any problems. And symlinks with my openfiler NAS worked without issues (the DIR in question is a slackware-13.0 mirror, made with a linux mirror script and which obviously contains links). I actually never suffered from the mounts being smb/cifs and performance with nfs was just 10percent higher, so I was happy with that solution and was trying to keep it that way when moving over to unRAID.

Thanks for the fast reply WeeboTech! Sorry for not being specific enough in the first place. When I mount the share via smb/cifs from a Gnome client and then try to copy over some directories from my openfiler, via that client I get the response "symlinks not supported by backend" for each symbolic link in the source-Dirwhen copying reaches the file. ...I can replay what you did, locally on my unRaid box though. So I would be happy to declare that symlinks are supported, if only I knew how to achieve the above from a client (linux-based) PC.

OK, my MoBo finally arrived...they are really hard to source at this part of the world. Here's a PIC with the MoBo mounted into a SC503-200B 1U case: I didn't have a PCIe card ready, but a PCI will show the details as well: The slot is approx 1inch apart from the card (ruler shows cm, not inches) Anyone to recommend a good flex-riser card for this? regards, fred edit: resized pics

Hi folks, I have now build my first unRAID box (on a plus lic.) and ran into an issue when trying to migrate my data. I have an openfiler NAS running which I want to retire as soon as the data has been migrated onto unRAID. Now I learned the hard way, that unRAID does not support symlinks Is there a recommended way to either - overcome the problem (i.e. while migrating - by automagically doubling the files instead of creating symlinks) - or to enable a symlink feature TIA, fred

yes, I owe you this, guiri...another problem solved by the helpdesk from hell, I'd say. On the other hand...well, I have to admit that there might be some good reasons why my brain archived that information somewhere in the dungeons. I am not going into detail, but...well, one knows that one has some "mixed emotions" when your mother-in-law, while taking a ride on your Harley, gets involved in a nasty train accident

...not as far as into production (with license)....I tested it with an unraid ISO instead of USB ...I'd also prefer KVM instead of XEN...USB access should work, PCI passtrough is also possible in newest KVM (but hardware needs to support it; intel vt-d extension, i think) and raw access to host devices is working out of the box. But you'd suffer from a performance degree, even if you inject virtio-block and -net drivers. I am currently running openfiler as a guest for my NAS. Average performance is about 60% than that of a native instance...your mileage may vary. I strongly recommend to give a look into ProxMox-VE. It is KVM on a small debian host with Web-based provisioning....rock stable and easy as 1-2-3 for KVM and openvz.

Guin, yes I understand and know what you mean. While I actually was considering your offer, it sparked me with the knowledge that a far-away relative is currently over in the US, working there on a green-card....I am getting senile ...

..speaking for myself, I don't... Hm, yes you're right...svnserve would/should also work...are you sure that the service is running and configured correctly?

I am not familiar with unmenu and the SVN packages installable from it ...SVN would need Apache incl. WebDAV in order to work remotely with TortoiseSVN. Is this working?...you should be able to see/read the repository using a webbrowser from outside of unRAID. I am not aware of a good HowTo on SVN and Slackware...there are some for other distros though on HowToForge....just in case you have some homework to do.

This sparks another thought on my side...this is what i observed from my stock unRAID in a VM: The hostname (not FQDN) is propagated into /etc/hosts but for the loopback interface. For eth0 (on DHCP) the hostname and FQDN is propagated back to DNS (and not into /etc/hosts) So a ping to "tower" locally in an unraid shell will resolve to loopback not the real net interface. -> every external client can resolve fine, but a local app could end up in resolving the IP wrong (incoming through an external trigger though, but what if resolve-calls are internally nested somehow?). Giving the IP instead of hostname would make this behaviour disappear. So...is it a bug or feature? unRAID without a real net makes no sense. Giving the hostname to loopback in /etc/hosts is considered as bad manners in some parts of the world I'd vote for a bug

Virtualbox is quite fine to that respect....I had good results with DVDFab under virtualbox. You can activate a pass-trough option for the drive. DVD Decrypter works as well. Choosing the right layer API for the drive (inside Win...is it called ASPI or NTnative or something?) seems to be of importance. What I like with virtualbox is, that it can run in headless mode too. And it offers 3D Support with their guest-extensions installed (even XBMC will work inside a guest just fine). Also USB support is superb, I'd say....a couple of days ago I had to do some team lesson on an external location ...and guess what?...the &§$"$§ LCD projector provided at the location came with USB only ...so I fired up my VBox-ed WinXP and connected it to my linux host laptop...I was able to install the WinXP drivers (the projector mimicked a graphics card and a CDROM with drivers), configured it and everything was up and running in 5mins. ...and we had some fun after hours, because the projector was a fullHD and I carried some decentral backups