thejonshow

Does unRAID have any known connectivity or performance issues once you get below 6% or 7% total remaining free space on the system? Just wondering if it's something as simple as that that's taking us offline.

Hey guys, I've had my massive unRAID setup (max'd out 24 bay server w/ 3 TB drives) for almost two years now running smoothly, thanks in large part to this forum. Recently things have started to really drag, dropping my read/write speeds to comical levels and even making the act of connecting to the server via AFP (our office is mostly Macs) via Finder a hit & miss process. Usually takes 4-5 login attempts from a user before the box will mount even with all green lights & acceptable temperatures across the GUI. The only variables besides age are that we've slowly been adding 1-2 users/month (now up to about 18 concurrent users), slowly running out of overall space, and we've added a totally separate Raid 5+0 to our network in addition to the unRAID box, but these servers are only connected via a switch. Some server specs below, along with a recent syslog file (broken up into 2 files). Any ideas for helping diagnose the apparent slow death of my server? Have I crossed a critical remaining space barrier that is causing the parity to drag or something? Running unRAID 5.0 Case: NORCO RPC-4224 4U Rackmount Server Case MB: SUPERMICRO MBD-X9SCM-F-O LGA 1155 Intel C204 Micro ATX Intel Xeon E3 Server Motherboard CPU: Intel Core i3-2120 Sandy Bridge 3.3GHz LGA 1155 65W Dual-Core Desktop Processor Power: CORSAIR TX650 650W ATX12V/EPS12V 80 PLUS BRONZE Active PFC High Performance Power Supply Controller: SUPERMICRO AOC-SASLP-MV8 PCI-Express x4 Low Profile SAS RAID Controller Memory: 2x Kingston ValueRAM 4GB 240-Pin DDR3 SDRAM ECC Unbuffered DDR3 1333 Many thanks. Logs.zip

Thanks guys.

Does this still apply in the case of an accidental unclean shutdown? The parity check auto-starts when I restart the array.

Got it. Could you in theory cancel a parity check to improve speeds temporarily, then run the full parity check later on? Or would that introduce possible parity corruptions since you'd be writing to the server temporarily without a valid party in place?

What is the impact of read/writing to a server during a parity check? Is there any issue or risk of data or parity corruption for writing to the array before the check completes? Or is read/write speeds just slower during this process? Thanks for indulging the novice question.

Hey guys- Realizing this is well outside the usual unRAID setup/support discussions, but since my unRAID box came up as a potential cause in this issue, I thought I'd kick this out to the forum to see if anyone had any advice or insight. I, like many of you, am the de facto IT department at my small business. In our previous location, I had run all the Cat6 cabling myself to various computers/switches, and we never really experienced any connection issues other than the internet randomly going out. Recently we moved to a bigger office that already had Cat6 cabling professionally run throughout the office to wall jacks that had been installed in each room. Convenient right? Well, out of a wired network of about 15 machines (all late-model iMacs), we all intermittently lose internet signal at least 3-4 times a day now, and not necessarily at the same time. The confusing thing is that the internet itself isn't going out, merely our connection to it. If I enable/disable the WiFi adapter on each machine, the internet roars back to the machine. But, whether you leave the machine wired or on WiFi, within a few hours the internet will slow to a crawl/disappear, and you have to unplug/replug your cable or enable/disable your internet adapter to get back online again. Some more details: - Office network is exclusively late-model Apple iMacs/Mac Pro towers. - We have a combo uBee router/modem from Charter Business Internet. - Running modem signal into a 24 port Unmanaged 10/100/1000 Mbps Gigabit Ethernet Switch from Monoprice. - Some offices split the signal from a wall jack through a NETGEAR 5 Port Gigabit Business-Class Desktop Switch to multiple machines. - The connection issue doesn't seem to be happening office-wide, but intermittently drops out on different people's machines at different times, but I can't 100% confirm this as we're all seldomly surfing the internet at the same time. - We do have a 24 bay internal unRAID network server connected to 1 of the 24 ports on the Monoprice switch for network storage. - I have bypassed the Monoprice switch and hard-wired directly into the modem and experienced the same connectivity drop-out (although the modem was also still connected to the Monoprice switch at the time). Charter has been out to investigate the issue 3 times so far, and they can't spot anything wrong with the outside lines or the modem itself (they've replaced it once already). All of their external diagnostics come back clean, which leads me to obviously assume this is an internal issue. One of the repair guys mentioned that our server or one of the devices on our network might be corrupted and causing some kind of IP flood or something (which is why resetting IPs seems to address the issue temporarily on each device), but that's something I'm totally unfamiliar with. Any advice from you guys on how to diagnose this issue further? Thanks for reading this far.

Wow. My server motherboard has 3 ethernet ports, two of them return no signal, and the last one worked. unRAID server restored!! Very odd, but thankful nonetheless for everyone's responses. Can sleep better tonight for sure.

Hey guys... so this is probably the issue.. the server doesn't ping. Ping results from 8.8.8.8: http://imgur.com/etqMOap Report from ethtool eth0: http://imgur.com/cEqsitO IP/Mask/Gateway of the machine I'm trying to access the server console from: http://imgur.com/9ep6kJd When I plug the cat6 cable feeding the unRAID server directly into a machine (iMac), there are no connectivity issues after disabling WiFi, so that connection to that router port should be fine. Any thoughts? Thanks for the insight.

ifconfig results: eth0 Link encap:Ethernet inet addr:192.168.44.99 Bcast:192.168.44.255 Mask:255.255.255.0 lo Link encap:Local loopback inet addr:127.0.0.1 Mask:255.0.0.0 UP LOOPBACK RUNNING MTU:65536 Metric:1 some other details about RX & TX packets, but no errors/drops.

Wow, very helpful breakdown Steve. After reading your post and doing some further investigating I was able to pull off what I learned to be a dustcover off that extreme 6+ wall connection and discovered it was 568B after all. The other thing that threw me, as you figured out, was that the existing starting points must have been wired to a serial terminal and not individual jacks previously. My unfamiliarity with both left me stuck in the mud. unRAID forums to the rescue once again. Many thanks-

Many thanks for your reply Dale. So the GATEWAY address in my network.cfg is not the GATEWAY address written on my router box from my new ISP? They left me with some new numbers for IP, Gateway & Subnet (all different obviously from my old location)- but I'm guessing those aren't important to unRAID? When I look in the network settings of other computers using the new network wirelessly without issue, I see that they have a a Router address of 192.168.44.1 (which, from my googling, I guess "Router" means "Gateway" on Mac OS?) Here's the network settings from one of those computers already using the new network: This is what's currently in my network.cfg file: # Generated network settings USE_DHCP=no IPADDR=192.168.44.99 NETMASK=255.255.255.0 GATEWAY=192.168.44.1 ... and when I type 192.168.44.99 into a browser, nothing happens.

Recently I moved offices and took my unRAID box to a new location, where I have a new router. Now trying to get the system back up & running. My understanding was moving the server to a new network would only require me to update 3 things in the network.cfg file: 1) IP 2) Netmask 3) Gateway Am I wrong? The router at my new location obviously has different external IP, gateway & subnet addresses that I would think I need to update. I tried using those and every time I try to go to the IP from my browser the address times out. I went into the network settings of some local Macs also on this new network, saw they were using DHCP and had their own 192.168 addresses, so I tried to use those too (with a different unique identifier at the end)... still no dice. Can anyone help me further diagnose this issue? Would love to get my unRAID box back up & running

Hey guys- Moving my unRAID box over to a new office that has Cat6 cable run neatly into wall jacks in all the rooms. My problem is I've never seen the layout & pinout configuration I've stumbled across in this new office, and I'm unsure what pinout config to use at my starting point to feed all the the rooms correctly. I have these as my termination points: and ... which appear to be some kind of Gigamax eXtreme 6+ connection I'm unfamiliar with. The pinout it's using looks really bizarre to me. Then I have these as my starting points: and which appear to be from a pinout of: Light Blue > Blue > Light Orange > Orange > Light Green > Green > Light Brown > Brown.... which I've NEVER seen before. Am I safe to rewire my starting point cables to a T568A or T568B connector with these existing termination points? Would one be more beneficial than the other? Any advice would be greatly appreciated. -Jon

Incredibly helpful gary- thanks. For the newbies that may stumble across this thread in a future forum search, if you don't want to hit a power switch to shut it down, you can run the following command after following all the steps of the wiki. shutdown -h now