before you say anything: i know that transporting an unRaid server inside of a huge check-in luggage is less than ideal to say the least, especially on a Munich - San Francisco international flight, but time restraints made it necessary.

 

with that said, the server (in one of those elegant aluminum Lian-Li cases) had a dent on the side of the top panel, which immediately made me apprehensive...it was sandwiched in about 3" of mattress-foam on all sides, so it must have gotten quite a hit to have a dent like that, but after opening it and checking for any loose components, everything seemed snug and in place...it actually booted up fine, but couldn't be reached at it's IP number from my iMac for the first 15 or so minutes after booting up...then, finally, i saw its icons pop up in my finder, and all the shares as i had left them before transport...but while clicking my way through some of the shares, it all of a sudden disappeared from the network again.

 

after booting up my MacBook Pro a few minutes later and typing in the unRaid IP into my MBP's browser, the web admin panel of my unRaid would come up again...but a few minutes later it would again disappear...it seems as if it disappears the minute you start interacting with it.

 

because of this behavior, i've been unable so far to generate and copy a system log from it...i will continue to try and post it here once successful, but in the meantime, does this type of behavior ring a bell with anyone? my gut feeling says something along the lines of "bad motherboard"...but i hope it ends up being something more along the lines of "push in the this or that a bit harder".

 

i'll be back if i can get the syslog.

 

Update 07:03pm: by unplugging and re-plugging the network cable to the unRaid, i managed to get in just long enough to generate the attached syslog.

 

Update 07-24-2012: after 24hrs of no glitches, the unRaid server being available and behaving normally, i'm going to call this one "Solved"...probably a bad cable, or one that was too "slow" for my Gb network (it might have been a Cat-5 cable, where a Cat-6 was necessary).

syslog_2012-07-20_0646.txt

since i don't really know how to make any sense of the syslog, i am dependent on someone more seasoned than me to find any hints as to why the server would drop in and out of the network.

 

i wonder how *just* the network functionality of the mobo could be broken...of course, maybe it's more than that, but i wouldn't know how to diagnose that...has anyone of you ever encountered just one chipset/function on a mobo fail? and if the hard bump the server seemed to have received broke something, wouldn't it be more likely that something got knocked loose, rather than breaking a mobo component?

 

considering that there is still one short PCIe slot available (is that a PCIe 1x slot?), would it be conceivable to pop another network card into it, to take over for the faulty one on the mobo? if it's worth a try, does anyone know of an inexpensive but good Gb network card that's supported by the kernel out of the box? (Amazon link, if possible?)

 

i'm dead in the water with this right now, and don't just wanna drop $150 on a new mobo, then hours and hours of work swapping it, only to find that it wasn't the mobo afterall.

 

looking toward you fine folks to give me a nudge in one direction or the other.

You could try some basic system tests; like a Memtest for 12-24hrs (that wills tress your CPU a little and memory - generally a good place to start to see if your motherboard was damaged.)

 

Your idea of trying another network card could work.  It appears you have a Realtek network card that may be the issue; what version of unraid are you using?

I would power off.  Pull and reseat the memory.  Pull and reconnect every disk and power connector in turn.  Check that the CPU heatsink and fan are properly secure.  it could be the CPU is overheating and running slow. And yes, run Memtest for a good number of passes as well.

i'll be back if i can get the syslog.

 

Update 07:03pm: by unplugging and re-plugging the network cable to the unRaid, i managed to get in just long enough to generate the attached syslog.

It seems you have a static IP address configured.

Jul 20 17:03:07 unRAID logger: /etc/rc.d/rc.inet1:  /sbin/ifconfig eth0 hw ether 00:30:48:B0:A5:A0

Jul 20 17:03:07 unRAID logger: /etc/rc.d/rc.inet1:  /sbin/ifconfig eth0 192.168.2.201 broadcast 192.168.2.255 netmask 255.255.255.

Is it possible you have a conflict with another device on your LAN with the same IP address assigned?  That might cause an issue similar to what you are experiencing.

thanks for all the suggestions!

 

@marcusone: how would i go about running a memtest? i doubt that i would be able to have access to the unRaid for 12-24hrs in its current condition...if memtest writes a logfile to the Flash stick, it would probably be ok if it dropped off the network for most of the testing period, but still wrote the final result back to the USB Flash stick...to answer your question, i've been running unRaid v.5.0-rc3 for weeks without problem before this transport.

 

@S80_UK: i powered the machine down, opened 'er up, and did what you asked...i found one power connector on the back of the middle 5-in-3 enclosure to be loose, but even after re-seating it, the fan wouldn't spin up (note to myself: get new fan for one of the 5-in3's)...all other cables, cards and plugs seemed to be firmly seated...i also loosened and re-seated the memory modules for good measure, as per your suggestion....as soon as i find out how to run a memtest on the box (i am pretty inept with the terminal), i will surely do that.

 

@Joe L.: yes, i've been using that IP number since i set the machine up over 5 years ago...i just did a first quick check on my other IP-enabled devices, and none of them used that same IP...there are two more people in this network, and once they are awake, i will check their devices as well to make sure, but i doubt that they grabbed that IP...not a bad theory, though...i wish it was something as simple as that!

 

FYI: i am using this Supermicro CSEA-O board: http://www.supermicro.com/products/motherboard/Core/G45/C2SEA.cfm

 

if anyone could take a quick look at these NIC's on Amazon, which one of them could i drop into the remaining PCI 1x slots to test whether it might only be the on-board NIC (which i still find to be unlikely, but if i can find a card for around $20, it would be worth the experiment:

http://www.amazon.com/s/ref=nb_sb_ss_i_5_4?url=search-alias%3Daps&field-keywords=pcie+network+card&sprefix=PCIe%2Caps%2C325

would it be a matter of inserting it, plugging in my network cable into the new port and that's it, or how would i go about disabling the onboard NIC, assgning an IP number to the new one, and using *only* the newly inserted NIC?

Select memtest86 during boot. This requires a terminal connected to the server.

 

Try a new LAN cable.

dang, i'll have to get a hold of an LCD monitor and PC keyboard before i can do that...i can't even do a clean power down right now, as i can't seem to be able to connect to the web-interface even after un- and re-plugging the ethernet cable a few times...with unRaid 5.0rc3, hitting the power button on the server once will *not* initiate a clean power-down, right?

 

also, if i remember correctly, unplugging the network cable from the back of my unRaid used to trigger a short "beep" coming from the server's mobo, or do i remember that incorrectly? because it stays silent now when i unplug the network cable.

 

i think my best bet (and certainly one of the least expensive next steps) would be to try a new NIC...again, anyone have any suggestion as to which of these ones would be plug-and-play for my unRaid 5.0rc3:

http://www.amazon.com/s/ref=nb_sb_ss_i_5_4?url=search-alias%3Daps&field-keywords=pcie+network+card&sprefix=PCIe%2Caps%2C325

 

i'll also try to find a way to borrow a screen and keyboard to be able to interface with the unRaid box directly.

Update 10:55am: I was able to gain access long enough to do a clean power-down.

i still don't have an LCD monitor to check the unRaid's output directly, but i heeded first dgaschk's advice of using a new network cable (which appears to have stabilized the dropping on/off the network, and then Joe L.'s advice of checking into the static IP address, and changed it to another arbitrary (but easy to remember) IP address.

 

the unRaid server has been running for about 10mins now without dropping off the network (the longest since the transport), and i'm keeping my fingers crossed that it was something as simple as a network cable (which i've been using with another machine without problems for months...curious) or an occupied IP number (since one of the people that used a computer here left yesterday, maybe they did have the same one as my server).

 

i'll keep monitoring it's behavior and will report back...here's to hoping!

 

Update 12:15pm: unfortunately i spoke too soon...shortly after posting the above, the server started to drop off the network again, and i haven't been able to reach it since...in order to exclude the possibility of another computer with the same IP number interfering, i did a direct link between my MacBook Pro and the server, with nothing else in-between...shouldn't i be able to bring up the web-interface with a setup like that, just by typing in the IP number of the unRaid system? i don't seem to be able to...my MBP is at: 192.168.2.3 and i had the unRaid system back on 192.168.2.201 before it disappeared again...subnet is 255.255.255.0 on both machines...if i ping the server IP address from my MBP i get a return time of less than 1ms...but no connectivity to the web-UI...i'm confused.

after a bunch more trouble-shooting steps, i seem to have found the culprit in a combination of cable and ethernet-hub issues...the cables i last purchased were Cat-5e, and my network and hub all run on 1Gb ethernet...apparently some of my cables don't work at all, some will work intermittently, and a couple seem to work ok.

 

for safe measure, i changed the IP number back to the new 192.168.2.55 (just to heed Joe L.'s advice) and put the one cable that's been known to work on my other machines on the server, and so far it has been up and reachable for over an hour.

 

i'll keep it running for another day or two before putting a "Solved" label on this topic, but for the moment it appears that a combination of cheap/wrong cables and maybe a little bit too cheap of a Gb ethernet-hub (it's a TRENDnet TEG-S80g) are to blame.

 

will update this topic in a couple of days...in the meantime: thanks y'all for pointing me in the right direction!

 

after 24hrs of no glitches, the unRaid server being available and behaving normally, i'm going to call this one "Solved"...probably a bad cable, or one that was too "slow" for my Gb network (it might have been a Cat-5 cable, where a Cat-6 was necessary).

 

thanks everyone!