I am at the point of embarassment now. Guests come over, I try to play movies from unraid, but the network connection is down. I reboot, and disk 11 is missing. Reboot again, no results, have to use poweroff before the disk reappears, and by then it is too late, my guests have moved on...

 

My motherboard has only to PCI slots, and they are taken by the expansion cards. I can ony add a PCI-X gigabyte card, and tried that, but the unraid server won't start, and limetech has not given me any support or explanation on my problem yet.

 

I am now thinking of starting from scratch again, and getting a new motherboard.

Ideally I am looking for a full size motherboard, with plenty PCI slots (in case I need them for a gigabyte card) and as many SATA connections as possible.

 

What are you guys using at the moment, that works well (no disconnects) with the latest version of Unraid?

 

Please throw a line to a drowning man....

What mobo and network chipset do you have?

The Mboard I am usin is the Intel Desktop microATX Motherboard D915GAGLK.

 

It was the motherboard used by Limetech in the first unraid build, so there are no software issues really.

 

I found the ABIT IP35 Pro browsing the web, seems like a good motherboard.

 

http://www.abit.com.tw/page/en/motherboard/motherboard_detail.php?pMODEL_NAME=IP35+Pro&fMTYPE=LGA775

 

8 SATA ports on the motherboard and 3 PCI (which I need in case the network chip is not supported).

 

Has anyone used this for Unraid? Will it support it?

 

Regards

 

I am using the Abit AB9 Pro. Everything works in version 4.0 But with 4.2.1 the Jmicron controller doesnt work properly and the on-board realtek NIC also doesnt function correctly. I am waiting for Tom to fix these two issues. I am recieving an Intel PCI-e gigabit network card totday. I'll let you know how it goes.

Would the onboard controller impact the SATA ports? I can use a separate network card but with 11 drives already I need to use the on-board sata ports....

Thanks for your feedback mate.

No help from Limetech, my server is down again.

Not really impressed with the level of customer service Limetech I must say.

I decided to replace hardware to find a solution, as at the moment my unraid server is unusable.

Bought a new motherboard, the GA-G33-DS3R, and already have supported network card which should work with this motherboard.

Will report the results.

You have a couple threads going & looks like you've replaced h/w?  Let's close off those threads and continue here, ok?

 

So... what is the current state of your h/w and what is the issue?

Have not replaced any hardware yet.

 

History:

- bought a PCI network card, but realised that there are no PCI slot left (now sitting on a shelf)

- bought a PCI-X card, but disks would not mount (now sitting on a shelf)

- bought a new motherboard with plenty of PCI slots and 8 onboard sata ports (awaiting delivery)

 

In the meantime, my unraid server keeps disconnecting.

 

I go to the server, type reboot, and voila, the server reboots and reconnects to the network.

Until I need it again. When I do play a movie, a week later or so, the server has disconnected again, or disconnects whilst I am playing a movie.

 

I already attached a log file, let me know if you need another one.

Is this something that has always happened with your system, or just recently?  How often does this happen, and can you make it happen?

 

There are many reports on the linux mailing lists of sky2 driver problems (though using older kernels than unRAID OS uses).  One suggested workaround is to edit the file 'syslinux.cfg' and change this line:

 

append initrd=bzroot rootdelay=10

 

to

 

append initrd=bzroot rootdelay=10 disable_msi=1

 

IMHO though, it's probably a marginal h/w problem with your motherboard.

 

FWIW, instead of rebooting, you should be able to type this at the console to bring the network back up:

 

rmmod sky2
modprobe sky2

 

I too have a lot of h/w sitting on shelves. 

 

 

My unraid was disconnected today too, I typed

 

rmmod sky2

modprobe sky2

 

and low and behold it works now.

 

I cannot replicate it, there is nothing I do that causes it, it just does sometimes lose its connection.

 

I have had disconnects always, but never every week, which is what is happening to me in the last two months.

 

I updated the syslinux.cfg as suggested.

 

Do I need to reboot for it to take effect?

 

 

Do I need to reboot for it to take effect?

Yes.

Rebooted, let's see if the disconnects continue.

And low and behold, it has happened again... barely a week later.

I will upgrade the motherboard soon. Got the new one but too tired to do it right now.

although I'm still running 3.xx on the original Intel board, I had similar problems long ago... in trying to diagnose the problem, at one point I removed one of my two 512mb sticks and lo-and -behold, running on one stick at 512 mb cured the problem!  I have no clue as to why, but it's been running that way 24x7 for about 10 months

 

Time to upgrade the motherboard.

I am going to uninstall the processor from the current motherboard and put it on the new one.

Do I have to scrape the thermal paste from the old heatsink or will it work as it is?

Time to upgrade the motherboard.

I am going to uninstall the processor from the current motherboard and put it on the new one.

Do I have to scrape the thermal paste from the old heatsink or will it work as it is?

 

Emphatically yes!  It has to be scraped, thoroughly cleaned, and fresh compound applied.  Check out the Arctic Silver info here http://www.arcticsilver.com/index.html, especially the product ArctiClean http://www.arcticsilver.com/arcticlean.htm.  Arctic Silver is a highly recommended thermal compound for CPU's, and ArctiClean is one of the best (perhaps IS the best) cleaner and prep for CPU and heatsink surfaces.  You don't have to use these products specifically, there are others, but the info on those links is very useful, to understand the importance of proper installation of thermal compounds.  Poorly cleaned and applied thermal compounds can cause the CPU to very rapidly overheat and fail.

 

Every household needs to have a can of GoofOff in the junk-drawer.... notonly does it quickly clean up heatsink compound, it removes gummy residue from labels and tape, bugs from the car and Soooooo many other things..... 

 

Time to upgrade the motherboard.

I am going to uninstall the processor from the current motherboard and put it on the new one.

Do I have to scrape the thermal paste from the old heatsink or will it work as it is?

 

Clean off all the old thermal paste using 70% (or better) isopropyl alcohol.  It is extremely important to have a solid thermal bond between heatsink and CPU.  Google to confirm.

 

 

Bill

Usually you need to use new heatsink paste. 

 

Be careful to not use too much.  The Intel site I found instructions on said to use heatsink paste about the size of a grain of rice and then to spread it first with a piece of stiff paper before putting on the heat-sink.

 

Joe L.

thanks guys, all your advice has been gladly received, and I am putting it at good use now, as I have just finished cleaning heatsink and cpu.

I was about to smear my cpu with paste, a grain of rice will be instead....;-)

Upgraded succesfully to the GA-G33-DS3R.

Now using a single memory stick 512 MB DIMM2 (a user reported disconnects stopped when moved from 2 sticks to 1).

Now my system is also ready for 15 drives, currently have 10 500 GB Seagate Sata drives. I filled the 8 onboard slots and used 2 slots of the 1st expansion card with no problems.

Using 4.3 beta 1.

Only problem is that the board does not allow me to reboot by flicking the power switch at the back. Thankfully bios allowes me to use the "power on by keboard" function.

I type a password and the system boots from powered off. Not a biggie but maybe someone else has found a different solution.

Now let's see if the disconnect continue.... currently using the onboard Lan, all fine. I already have a PCI card ready and waiting if there is any need, and a free PCI slot.

I must say that using one DIMM2 stick (was running 512 before too, but two sticks DIMM) and this mboard unraid boots up twice as fast.

Touching wood, so far so good....

Well well, first morning, first problem...

Was playing a dvd from the server this morning, to test the efficiency, and all worked well.

When I then typed tower in the web browser, I got a glimpse of the management window, and of a drive (the second) with a red icon next to it.

Then the page changed to shutting down server.

Server rebooted and now there is a red dot (left side) next to drive 2.

I will attach a pic of the drive management page and the log file.

Be careful to not use too much.  The Intel site I found instructions on said to use heatsink paste about the size of a grain of rice and then to spread it first with a piece of stiff paper before putting on the heat-sink.

Joe L.

 

The instructions on artic silver say to let the heatsink itself spread it. Place heatsink on top of core, twist one/two degrees to right and left.

I.E. Use the size of a grain of rice, put the heatsink on. Press and twist it "slightly" to spread the compound underneath it.

 

http://www.arcticsilver.com/instructions.htm

 

Find attached a snapshot of my management window

Either drive 2 has failed, or the cable to it is loose... or far less likely, the SATA controller port failed.

 

The "reboot" when going to the management console has been reported before... So far, it is thought to be as a result of the browser caching the "reboot" command from a prior shutdown.

 

Joe L.