Spectrum

ooh this sounds like a perfect application for a greasemonkey script..... Of course I don't know javascript so I couldn't do it The other way I thought about doing it would be using a Perl script to match regexes and spit out HTML; that I could whip up in no time flat, ok a few minutes, but I would need a list of things to match against. I can scan a syslog and have pretty good feel for what is going on, but I don't have a comprehensive list of what should be what color.

I think I would start by troubleshooting what is causing the interrupts on the ESXi machine; that does NOT sound normal. That said, if it is something to do with pscp, you could try installing cygwin to get a bash shell and a "native" scp application that might behave better. Another line might be setting up the ssh client libs on the unRAID box and initiating the transfer from the unRAID box via a cron job or from another system using plink. I haven't tried getting openSSH up and running on unRAID so I can't say how hard this would be or how well it would work, but there is an unMENU package for it All in all I think you would be best served by figuring out what is causing the interrupts and solving that problem.

Anytime Networking can get really complicated really quick when you start mixing different paradigms (MS/Windows/Mac). The only thing they really all agree on 100% is the IP/TCP/UDP protocols.

Name resolution on your local network is not using DNS, it is using NETBIOS, these are two different protocols for mapping names to IP addresses. DNS is used for the Internet and in enterprise situations and is typically not worth setting up on a small, single subnet, home network. Most consumer routers act as a DNS proxy so whenever it receives a DNS request it forwards it to the ISP DNS servers; this is solely for resolution of public domains (.com .net .org etc). The problem is that the Supermicro IPMI modules do not use NETBIOS for resolution, so your options are set the IPMI device to a static IP address, set a DHCP reservation on your router/DHCP server so the IPMI device always receives the same IP address, or let the IPs fall where they may and use the Supermicro IPMI management utility and let it scan for the device.

Hah no one's even downloaded my mid-session syslog, oh well. It's been a month and I have a syslog captured for an L2 test of this board. The first parity check is on Nov 6 and the 2nd is on Dec 1. I can run another parity check if the monthly Dec 1 check is not deemed close enough to the end I don't see anything in the logs that look troublesome, but I would appreciate it if someone (Raj?) would confirm that this is good enough for considering this board Level 2 tested. System is: Motherboard/CPU: MBD-X7SPE-HF-O (Newegg Link) RAM: Kingston 2GB DDR2 667 (PC2 5300) Model KVR667D2S5/2G (Newegg Link) Power Supply: CORSAIR Builder Series CX430 (Newegg Link) Harddrives: 4x WD20EARS-00J, 1x ST31000523AS, 1x ST31000340AS Flash Drive: Lexar JD FireFly Nothing else is plugged into the system other than a KB/mouse. The pcie slot is empty. Edit: I haven't kept an exact count on transfer to/from the box during the past month, but it is >= 1TB for sure. Any questions or concerns, please let me know! MBD-X7SPE-HF-O-syslog-2010-12-07.zip

5 minutes seems short, but Linux will clear the screen after a timeout to prevent monitor burn in. Try pressing an arrow key on the keyboard and see if it comes back.

Seagate had a bad batch of 7200.11's or 7200.12's. Don't remember the actual models, I would have to crack a case to find out. Ended up going through a single manager that managed the incident from support and it took 5 drive shipments inside 2 weeks to get 2 drives that would not fail inside 5 hours of use. This is testing the drives in diff facilities under diff operating conditions. Neither system killed drives before or since. The manager would not out right say the words "bad batch" but he danced very close to it. As far as 2k drives, that was just a number pulled out of the air. I have no idea how many facilities the drive manufacturers use, what their production runs are, or what their QC standards are; but sh*t happens and sometimes it doesn't get caught until it's too late. Companies that deal with mass production understand the risks and what to do about it much better than we do. That's one of the reasons there are warranties, dealing with the faults that do get through the system. Yes it may hurt their rep for a short period of time if a company the size of Seagate/WD/Samsung sells a few thousand bad drives, but it would probably not affect their stock price by 1 full point, and that's what matters most to these companies; bottom line.

Generally a bad batch in this sense is when there is an entire manufacturing run with defective (or substandard) parts. ie. Seagate's manufacturing facility receives a batch of IC's that were contaminated during one of the many lithography steps at the foundry and fail pre-maturely due to heat. These IC's are used in the next 10,000 drives manufactured and Seagate now has a "bad batch" of drives that are prone to fail early in their life cycle. It has nothing to do with the design of the drive, or the design of the IC that was used. It fails based on a manufacturing error. While not common, this kind of thing happens. The idea is that there are quality checks along the way that should catch this, but they don't always work.

Reminds me of this

Or while browsing you come across an embedded image that is hosted at imageshack. You really don't realize how much content crap is downloaded from sites other than the one you are visiting until you install noscript

If those are from the outgoing log, it Looks like web traffic originating from whatever machine is at 192.168.1.101 on your internal network and the destination being the other IP address. www is standard web traffic on port 80, and https is encrypted web traffic on port 443. 174.36.153.36 resolves to nightingale.dotcolo.com 65.54.51.29 does not have a reverse dns entry but is in the 65.52.0.0 - 65.55.255.255 which is owned by Microsoft 208.94.3.72 is in the 208.94.0.0 - 208.94.3.255 block owned by Ezri Inc in Los Gatos CA but it's nameservers are listed as imageshack.us so I would assume they are affiliated. If 192.168.1.101 is the address of your unRAID box it might warrant a little more investigation, if a regular machine Windows/Mac/Linux has that IP, it's probably just web traffic.

Spaces are bad! Try changing it to The_Vault or TheVault. Does it really matter? Not if you want to address it by IP address than its name, but I find remembering and typing out IP addresses cumbersome. Also in between changes to names/IP addresses it is beneficial to run (on your windows box) ipconfig /flushdns nbtstat -R before you ping again to make sure your system queries the network for the current info. Don't want to get everything set right and not get a response because windows has latched onto a netbios name/IP. Once everything is configured, you don't have to worry about flushing the resolver cache or the netbios table.

Internal networking is independent of DNS resolution. Use whichever is more convenient

YW, glad you got it resolved!

Just go to the settings page, then set the following the Network settings panel Obtain IP Address Automatically = NO IP Address = 192.168.1.50 Netmask = 255.255.255.0 Gateway = 192.168.1.1 Obtain DNS Server Automatically = NO DNS Server 1 = 192.168.1.1 Then click apply. Stop the array, and reboot the server. When it comes back up log back into the web interface to make sure that is working. Log into the console and type ping google.com to make sure DNS resolution is working. If not, change the DNS Server 1 setting to 76.10.191.198, apply, reboot, try again. Once everything there is working, you can open a command prompt on your windows machine and type ipconfig /flushdns nbtstat -R ping SERVERNAME but change SERVERNAME to whatever you have your server named (default is tower) to test netbios name resolution. If you get replies name resolution works, if not see this article on the wiki for troubleshooting.

Just press CTRL-C a few times and it should dump you back to a prompt.

Your unRAID server and desktop are on 2 different subnets! Either you have VLANs set up with multiple DHCP servers, or you have your unRAID box set with a static IP to the wrong subnet. Log in to the console as root. Type cat /boot/config/network.cfg What I'm expecting is the result to be # Generated network settings USE_DHCP=no IPADDR=192.168.0.187 NETMASK=255.255.255.0 GATEWAY=192.168.0.1 If USE_DHCP is set to yes don't do any of the below, just post back that it is and we can go from there. To change the IP address on the fly type ifconfig eth0 192.168.1.50 netmask 255.255.255.0 then press Enter. I am assuming there is nothing else on your network with the IP address 192.168.1.50, if there is, pick another number and replace the 50 with it Now from your desktop put http://192.168.1.50 into the address box and press Enter. That should get you to the configuration page of the unRAID box. Use the instructions from one of my previous posts to reset the network settings to something appropriate to your network, then reboot the unRAID box. Everything should work when it comes back up.

If you have multiple devices behind the router and everything else is working you shouldn't have a bridging/routing issue with the modem. Odds are after a reboot, your router handed the unRAID box a new IP address. Check the IP address on the unRAID server using the console (method described in previous post) and see if it is different. If that is the case, after you get the new IP address, log into the web management and set a static IP address so it will not change on you. You need to get the network block and subnet from the DHCP settings on your modem/router (assuming it is the DHCP server) or another computer and set things up on the unRAID box accordingly. Just make sure you give it an IP address that is outside the range the DHCP server is using (this will be in the DHCP server configuration for whatever device is acting as your DHCP server, probably your DSL modem/router) or you could end up with > 1 device with the same IP which makes things stop working Once you get the IP stabilized you can look into troubleshooting why name resolution does not work http://lime-technology.com/wiki/index.php?title=Troubleshooting#Name_Resolution

Assuming your network is setup as Internet <--> DSL Router <--> Switch <--> Computers or Internet <--> DSL Modem <--> Router <--> Switch <--> Computers Use the keyboard and monitor attached to the unRAID box and log in as root. Once logged in type ifconfig then press enter. Look for the IP address associated with eth0 eth0 Link encap:Ethernet HWaddr 00:25:90:0a:50:66 inet addr:192.168.1.51 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1 RX packets:83193956 errors:0 dropped:0 overruns:0 frame:0 TX packets:80016046 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1556875105 (1.4 GiB) TX bytes:742174664 (707.7 MiB) Memory:fe9e0000-fea00000 If it is not something in the 192.168.x.y or 10.x.y.z range you have IP addressing problems. If it is in the aforementioned ranges, use that IP address in your web browser to connect by putting in that IP address as http://IP_ADDRESS (http://192.168.1.51) for the example above). If your network is set up as Internet <--> DSL Modem <--> Switch <--> Computers You should disconnect your unRAID box from the switch immediately as if it is getting an IP address it is on the internet and unRAID is in NO way secure enough to be facing the wild of the internet. If this is the case, your previous cable modem functioned as router and your current DSL modem is functioning as a bridge and you need a router in the mix.

MS actually has a relatively decent article here that addresses the TCP/IP basics that you should know for setting up a network. The gist for what you need is all computers have to be on the same IP subnet. Each device must have a unique IP address. Each device must have a default route to get out of the network (router to the internet). Each device must have the IP address(es) of DNS server(s) for name resolution. To access the network settings in UnRAID log into the web interface, click on settings, and make any changes you need in the Network Setttings panel.

I've had mine up and running solid for 2 weeks now for Level 2 test, syslog attached. Anyone see any bugaboos or things to worry about for Level 2 so far? syslog-2010-11-21.zip

Only way to run it headless would be via IPMI or KVM over IP. Also be aware that it affects ALL WD drives attached to the system; disconnect any WD drives you do not want changed before you run it.

This got me curious about my WD20EARS. for f in sda sdb sdg sde; do echo $f; smartctl -a -d ata /dev/$f|grep Load; done sda 193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always - 527 sdb 193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always - 513 sdg 193 Load_Cycle_Count 0x0032 199 199 000 Old_age Always - 3213 sde 193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always - 404 Can you guess which one is the parity drive? These drive were all purchased at the same time and are < 6 weeks old. Makes me think I should run wdidle on /dev/sdg but leave the others alone. Thoughts?

Yeah i hear that but I'm glad I'm not paying that power bill

Good to hear The hard drive that those files lived on when that happened finally gave up and died and the other guts are actually in my HTPC now. I memtest and prime95 test for at least 72 hours before I consider a new box "stable" so I'm really not that concerned with memory/cpu corruption, more if the underlying drive starts going out. Not to say mem/cpu won't go south, but I can't spend *all* my time worrying, I have too many things to do!