bblue

I'm sure either of those would improve things in a standalone world (I'll try later when I have a free drive). But in terms of how they seem to work in unRAID, the times I getting with the existing test and SATA 3Gb/s interfaces are just shy of those numbers. Typically 110-120MB/s during a pre-clear (with no other disk activity). But that could be (in part) by limits in this motherboard architecture. I'm going to test that notion shortly. The current MB is Intel D915PBL with P4 3.4Ghz processor, ICH8 chipset and is limited to DDR2-800 RAM (a 400Mhz FSB). I just picked up a Gigabyte EP43-UD3L MB which can handle P4 to quad core processors, and up to DDR2-1333 RAM (a 667 Mhz FSB), and ICH10/P43 chipset. With some fast ram and an E8400 3.0Ghz dual core processor, it should be easy to determine the effect of the motherboard on disk i/o. Which brings me to a question I was going to post separately about, but maybe you know. I read somewhere on the forum or wiki that Linux doesn't use but one core of a processor so the notion of a dual core doesn't help except for better speed-stepping, lower power consumption, and faster FSB/RAM. But if hyperthreading is active on a single core processor, does Linux take advantage of that, or is it strictly one thread whether it's running on an HT enabled processor or not? On my Intel board I've always run it with HT disabled, assuming only one thread can be used. Is there any beef to be had under any condition of more than one thread available? --Bill

As specified, dd if=/dev/sda of=/dev/null count=8192000 I'd assume that count would be in 512 byte blocks which are native to the drive, for a total of 4,096K bytes read in.

Ok, have finished the testing of the last four drives on sil3132 PCI-e cards. It's the same rate whether or not they are connected through a port multiplier (sil3726). Motherboard SATA ports (ICH8) 3Gb/s: Parity SATA 0 2T 111 MB/s Disk1 SATA 1 1.5T 123 MB/s Disk2 SATA 2 1.5T 122 MB/s Disk3 SATA 3 1.5T 122 MB/s Syba PCI card w/SIL3124 3Gb/s chipset Cache SATA 0 500G 65 MB/s Disk4 SATA 2 1.5T 81 MB/s Disk5 SATA 3 1.5T 83 MB/s Disk6 SATA 1 2T 84 MB/s Syba PCI-e card w/SIL3132 3GB/s chipset Cache SATA 0 500G 72MB/s Disk4 SATA 2 1.5T 94MB/s Disk5 SATA 3 1.5T 122MB/s Disk6 SATA 1 2T 120MB/s In the third group, the reason for the difference in throughput between SATA 2 and SATA 3, both Seagate 1.5t's, is that SATA 2 is a Seagate 1.5T LP (5900 RPM) model number ending in 541AS. All others are the standard Seagate 1.5T's (7200 RPM) and a model number ending in 341AS. Not sure quite how I ended up with that LP but it wasn't intentional. The LP version also doesn't show a temperature reading in either unRaid management, or unMenu. When a parity calc is running, everything goes to hell. All the Syba controller rates drop to about 13.3 MB/s for drives in the array, and with a port multiplier, or about 21MB/s without the PM. The motherboard SATA drives are also down to around 21MB/s. The cache drive in my case (a much slower Seagate 500G drive) drops to 30MB/s. If you have a bunch of files on your cache drive, mover will take forever and undoutedly slow down the parity process as well. With parity and mover running, load average rises way up into the 5-6 range, but CPU utilization peaks at just 40%, so a faster processor (mine is 3.4Ghz) probably won't make much difference. It's all I/O bound. So, some questions for anyone who might have experience with all this: 1. Do the port multipliers do just basic switching between devices depending on demand, much like two devices on an IDE controller? I.E., it's one or the other active at a time, period. Or are there more smarts to it than that where device switching is more transparent? 2. My numbers seem to imply that you'd be better off with multiple PCI-e 3Gb/s cards and motherboard SATA ports (native) than PCI-e with PM's, but this could just be the behavior of my board. General thoughts about this? 3. How much impact on the I/O (realistically) would a faster FSB and RAM be on a newer motherboard, while not necessarily increasing CPU speed? This board is using DDR2-800 RAM and it's operating pretty much 1:1 with the capabilities of the MB. But newer ones can achieve up to double that without overclocking. It seems that if this 122MB/s figure is not really limited by the drive, the faster FSB would improve things. 4. Maybe in some way relevant: The shfs process invoked to mount /user0 *apparently* starts using a lot of CPU (20-40%) and quite a bit of memory when mover begins, but when mover is stopped, the usage doesn't drop back to normal. Even with Samba killed and nothing camping on any shares, you cannot umount /mnt/user0. However, about 10-15 minutes later there's this message in syslog: Sep 10 08:34:13 Bench kernel: shfs[31002]: segfault at 0 ip b7f421a3 sp bf86f02c error 4 in libc-2.7.so[b7ed0000+146000] and the high CPU thrashing stops. But of course /mnt/user0 isn't there until you recreate it with "/usr/local/sbin/shfs -cache 0 /mnt/user0 -o allow_other,attr_timeout=0,entry_timeout=0,negative_timeout=0". Then restart Samba and the shares reappear and everything continues on. Does anyone recognize what is happening with the crazy behavior of this process? It takes around a minute or two before the CPU and RAM usage start escalating after mover starts. (I'm currently using 4.5b6) --Bill

It sure looks like it would work, as long as you have an available PCI-e slot. Newegg has a good return policy too, if it doesn't. --Bill

Thanks Joe! You certainly are a master script-writer. I'm no stranger to awk, but the use of the pair of triple single quotes eludes me. What is actually happening there? --Bill

And for the first question, you should see either very low numbers in DROPPED or zero. High values indicate either a physical cabling or switch/hub problem, a config mismatch between the interfaces, or the receiving interface unable to keep up with the transmit speed of the sender, and unable to 'pace' it with internal flow control. High dropped numbers will significantly reduce your transfer speed, but because of the built-in error correction in the TCP part of TCP/IP you are unlikely to end up with corrupt files. It is the sign of a problem to be concerned with, however. --Bill

Any time you do an ifconfig and eth0 doesn't even show up, it means that the driver in unRAID for that interface cannot (or is not) talking to the interface hardware. --Bill

Ok, finally got all the Seagate 1.5T and 2T drives installed in one place for testing. Motherboard SATA ports (ICH8) 3Gb/s: Parity SATA 0 2T 111 MB/s Disk1 SATA 1 1.5T 123 MB/s Disk2 SATA 2 1.5T 122 MB/s Disk3 SATA 3 1.5T 122 MB/s Syba PCI card w/SIL3124 3Gb/s chipset Cache SATA 0 500G 65 MB/s Disk4 SATA 2 1.5T 81 MB/s Disk5 SATA 3 1.5T 83 MB/s Disk6 SATA 1 2T 84 MB/s So the Seagate 2T stacks up reasonably well, overall, specially for a 5900 RPM drive. These read only numbers don't really tell the whole tale, however. Is the lower throughput on the PCI controller due just to this motherboard (Intel D915PBL w/3.4Ghz P4, 4GB RAM), or PCI in general? I now have a couple of Syba PCI-e boards with the Sil3132 3Gb/s chipset for testing, but haven't done so yet. The Sil3132 is the 2 port version of the sil3124, both of which support port multipliers based on the Sil3726 chipset. With Parity checking and Pre-Clearing taking as much time as they do, I'd certainly rather have all the drives up in the 110-120MB/s range than in the 80's, but need to find out why the lower rates on the PCI. Anyone with ideas about this? BTW, the boards I have tested with Sil3114 and Sil3512 chipsets, which are 1.5Gb/s SATA I, perform about 20% lower on this motherboard for the same drives. They are also PCI. --Bill

When you go into the file browser in unMenu, select a filesystem, and then click on a directory whose name has spaces, you get this response: File Browser Directory: invalid-directory/ Sorry: Access to invalid-directory/ is not permitted. Single name directories in the same filesystem work fine. --Bill

Hi Joe, I'm wondering why you disallow directories with spaces in their name in this plug-in? I know it's non-standard in the Unix world for file and directory names to have spaces, but it does work. I'm not sure I understand why it can't or shouldn't be used as a convention in exported movie/song (etc) titles? On unRAID they seem to export ok and clients don't seem to have a problem with them... Are there other areas in unRAID in general where this is not a good idea, or doesn't work? --Bill

Right, you should be able to with the IP address. Let's try something else. From your tower console, can you successfully ping the ip address of the windows machine? 'ping 192.168.0.101' If so, it will show you receiving 64 bytes from the other end every second. From your windows machine 'ping 192.168.0.102' in a DOS command window. It will respond similarly but stops after 4 cycles, if there is basic connectivity. Also, do an 'ipconfig' and let me know what it says. --Bill

Joe, Many thanks! Works like a champ. --Bill

I'm not finding that ssmtp package on the Slackware site anywhere from build 11 to 13. The only reference is to a libcddb that has a simple smtp backend. Any idea where to locate it? That version of mailx is there, however. --Bill

Very clever! That's what I first thought, but the 'signature' bytes threw me away from that notion for some reason. So what makes it show as uncleared after an unRAID format is that the filesystem type (83) is now defined, which breaks the tests. That seems like a lot of redundancy since bashmail can already talk to a local smtp server, it just doesn't have the right front end, plus relies on unraid-notify's conf file for default smtp info. Seems like it wouldn't be too difficult to make a new front end for bashmail and add a .conf file and reader just for it for a quick and simple delivery backend. I don't have a netcat, what's that about? Yeah... I guess that would be easier (less time involved) and more universal overall. I was hesitant to be loading packages that are far more than they need to be to get the job done. But then there's the time issue (mine)... --Bill

Just look at both end connectors. You'll see 8 copper connections, and a colored wire inside the plastic going to them. If there are 8 wires, you're ok (assuming the cable is actually good). If there are only four connected to pins 1,2,3 and 6 on either end it won't work on Gigabyte ethernet. Only 100Mb and lower. For example, when you connect to your ISP there is either an automatic or manually set group of parameters that include the ip address of at least one nameserver. The nameserver's job is to convert a name, like lime-technology.com to a corresponding ip address so that you can connect to it. Only registered domains and hosts worldwide are in this database. So if you wanted to add a name that represented an ip address that didn't appear in nameservice, in windows you would add the name/ip address pair in the lmhosts file so your system would know what address you meant when that name was used. Likewise, computers on a local network that don't have a local nameserver to resolve names must have entries in lmhosts (windows) or /etc/hosts (unix) to resolve to an ip address. Any time you put in an http:// or ftp:// type name you would be using either nameservice or local lookup. If you don't have this, you can only refer to the remote connection by address, not name. Note that the special syntax used for the initial connection to an unRAID server, //tower, does not use nameservice as it is a NETBIOS name, not an internet style name. --Bill

No, I installed the lan card, disabled the onboard, and booted into unraid. Then ifconfig did not show the eth0 details. BTW, that lan card works fine in windows. Its a realtek 8168c. The mobo is an intel dg965. Here are the listed compatible NICs. I don't know if it's complete, but I do know that some cards listed don't work for other, and I've found at least one that works but isn't listed. Network Controllers * Atheros L1 * Broadcomm BCM5751 Gigabit PCIe x1 * D-Link DGE-528T Gigabit Ethernet (from v4.1) * Intel 82566DC Gigabit LAN chipset * Intel PRO/1000 Gigabit Ethernet * Intel PRO/100 Ethernet * Linksys LNE100TX * Marvell Yukon Gigabit Ethernet * Netgear GA311 Gigabit Ethernet * Realtek - RTL8169S, RTL8111B, RTL8111C, probably others too * Others that use the same chipsets as the controllers above From http://lime-technology.com/wiki/index.php/Hardware_Compatibility#Network_Controllers --Bill

Thanks for all the info, Joe! When you say that unRAID "completely recreates it" (the MBR) are you talking about just when you put in an uncleared disk, or could that also happen during the formatting phase (pre or post)? I notice in your script you don't ever clear it, but just write to the appropriate places to size the filesystem, partition #1 and pointers. Is it just unnecessary as other bytes don't matter, or what? It would be ok if unRAID clears all of 1-62 sectors with any flag bytes in it, because I think I'd only want that extra flag there if preclear_disk actually did the work. Bytes 510 and 511 of MBR are preclear_disk.sh's flag, not really a function of MBR bytes, correct? That's the only place in your code I can see a signature write, even though it's labeled as signing the MBR. Are the 0x55AA values arbitrary or significant in some way (other than being mirrored alternating bit-sets)? On a different subject, which mail binary is supposed to be used with this script? It looks like it's the old standard /bin/mail by the arguments, but I don't find anything like that on the distribution site. Bashmail is a totally different animal although I found some apparently incorrect references on the Forum that if you had unraid_notify running you could use its back-end for mail. Oh man. It sounds like that would have been both exciting and incredibly frustrating at the same time! That was back when, during the early transition from tubes? Sounds like you were quite 'inside' at the hardware level. What insight that would give you! And who even remembers CP/M for that matter. It was kind of cool in its own way, though. At least compared to what else was available at the time. The whole OS could relocate anywhere in memory to suit the hardware or other requirements. I had (among other things) one of those 'portable' Kaypro 10 CP/M computers, maybe around 1982-1983. Very cool at the time with that huge 10M drive... --Bill

This sounds like you had the on-board NIC enabled at the same time using the card? You can't have two active NIC's. Try removing your NIC card and turning on the onboard NIC, see if that works. If the onboard is a Marvel chipset it might. Whether or not it works on the same hardware in Windows isn't relevant (well, it rules out dead hardware I guess), since with unRAID you are running a different operating system and different device drivers. Apples and Oranges. You need to find a NIC that works in unRAID. Look at the hardware Wiki for probable NIC's. Is yours on that list? (card or chipset). You'll need to obtain one that is and try it to see. But definitely keep the MB NIC off, and for that matter, serial and parallel ports too if not needed. It's a PIA to be sure. I tried a total of six cards and found one that worked if a SATA disk controller was NOT installed, and one other that seemed to work regardless. That one was a PCI-e card D-Link DGE-560T with a Marvell chipset. All the rest were PCI and played dead, even though they all contained what was supposed to be a compatible Realtek 8169S chip. Go figure. --Bill

As long as I use the motherboard's SATA ports (ICH8 on this D915PDL Intel board) I get 126 MB/s on the Seagate 1.5T's with that test. If I do the same on a budget PCI controller with a sil3114 or 3512 chip (1.5GB/s) it drops to around 90 MB/s. I have some better ones coming this week (with sil3132 and 3124 chips) and it will be interesting to see how they fare. Also, broke down and picked up a couple of the 2T Seagates (only $175 each at Fry's) so will try them in the next day or so and see how they do. --Bill

No. The router's basic function is to gain connectivity to the outside world, it can do nothing about a failed port on a system. You have to find an interface card that will work on your system and is correctly supported by unRAID. Check the hardware Wiki for possibilities. Though it's not likely, are there any BIOS settings in the motherboard having to do with ethernet connectivity or PCI ports that may be set incorrectly? If the MB has built-in ethernet, that should be turned off in the BIOS, etc. --Bill

What is the interface on the win7 set to? It needs to be an address in that same subnet (like 192.168.0.105), same netmask and no gateway set. If you're not using a hub or switch between the two machines you'll need an 8 wire crossover cable. (some generic cables only come with four wires connected through) If you don't have a nameserver on the windows side, you'll need to make sure the lmhosts file has an entry for tower in it, else you'll have to address it by ip only. For example, http://tower won't work without an lmhosts entry, but //tower should since it uses a different method of initial connect. --Bill

Yep, bad. The driver (if any) being selected for that card is either not there or not working correctly with the card. You'll see additional lines when it's actually working. Something like this: [Destination Gateway Genmask Flags MSS Window irtt IFace 10.146.162.0 0.0.0.0 255.255.255.0 U 0 0 0 eth0 127.0.0.0 0.0.0.0 255.0.0.0 U 0 0 0 lo 0.0.0.0 10.146.162.33 0.0.0.0 UG 0 0 0 eth0 The last line would only be present if you had set a default gateway for outbound traffic to a router or equivalent. However, you don't need that just for local machines talking to each other. Also, if you do an 'ifconfig eth0' with no arguments the second line will look something like this: inet addr:10.146.162.49 Bcast:10.146.162.255 Mask:255.255.255.0 But only if the interface is 'up' and working. I don't know about the chip set on the card you're asking about. It may or may not, so just give it a try. There seems to be something squirrely with some of the drivers included in the 4.4.2 build. --Bill

Destination host unreachable can mean a number of things, including driver problems in the unRAID computer. You should have no problem with your win7 computer, and your unRAID, both connected to the switch with a straight through cable. The route to the local net is implied when you assign two ip addresses from the same subnet, which you did correctly. In the unRAID box log into the console and type 'netstat -nr'. You should have three lines in response, if so, what are they? I'd bet you have 0 or 1 lines which means that the interface is down, usually because the drivers aren't talking to it. Look at the LED's for the ethernet connector, is one of them lit? Are there any lights when you 'ping' the win7's address? I'll bet not. I've run into several cards, for example, with the Realtek 8169S chip that don't work. One that does. Those that don't exhibit exactly the behavior you're describing. --Bill

Has anyone done any i/o or other performance tests on the Seagate 2T drive, compared to the performance of a Seagate 1.5T? All I've been able to find on it is one review that compares it favorably to a Seagate 320G (which is not a stellar performer) and some hints that it outperforms the WD 2T. But the WD drive scores pretty low on almost every i/o test, so that doesn't say much either. I'm trying to determine whether I'd take a big hit in performance by using the Seagate 2T for parity, and planning for other data 2T's when the prices come down some more. Any thoughts? --Bill

An additional option is in your go script is to install "xz", followed by an upgrade of package utils themselves [ie: ftp://slackware.mirrors.tds.net/pub/slackware/slackware-13.0/slackware/a/pkgtools-13.0-noarch-3.tgz ]. In the future you wont have to deal with converting the .txz files to .tgz anymore. I've put this sequence in my go file for now. Hopefully the build used for unRAID 4.5 will have this stuff built in. --Bill