Would you believe it, Microsoft have released the iSCSI software target for all users of server 2008 R2!
http://blogs.technet.com/b/josebda/a...-download.aspx
Apart from server core
On server core, it would make a nice less resource heavy storage box.
"In a perfect world... spammers would get caught, go to jail, and share a cell with many men who have enlarged their penises, taken Viagra and are looking for a new relationship."
How are people configuring Openfiler?
I'm thinking about GPT on 5 2TB disks, a hardware RAID 1 across all drives, just 10GB or so, then software RAID 6 across the remaining space.
Will that actually work? It seems that I need GPT on them to make it work, but I'm getting conflicting information about whether it will boot properly.
Also, has anyone come across the 95% capacity partition bug on here? There is some mention of it on the openfiler forums, but the advice doesn't get any better than "try v2.3"
Linux with GRUB 2 will boot from GPT disks fine even on a BIOS system.
jim (02-05-2011)
I don't think you can get any cheaper than this:
http://www.crescentelectronics.co.uk...in-ddr3-4.html
Brilliant price. I got it myself for my own server and it works great!
My Blog => http://adriank.org
Error correction is the major benefit but ECC memory cost a bomb...
ECC vs. non-ECC — What do I Have and Can I Mix?
When adding new memory, you should match what is already in your system. Adding non-ECC memory to an ECC system will disable the error-checking and correcting ability of your memory modules. While your system may still operate, the enhanced features of the ECC modules will no longer be active in your computer.
You can determine if your system has ECC by simply counting the number of black memory chips on each module. ECC (and parity) memory modules have a chip count divisible by three or five. This extra chip detects if the data was correctly read or written by the memory module. If the data wasn't properly written, the extra chip will correct it in many cases (depending on the type of error). Non-ECC (also called non-parity) modules do not have this error-detecting feature. Any chip count not divisible by three or five indicates a non-parity memory module.
Using ECC decreases your computer's performance by about 2 percent. Current technology DRAM is very stable, and memory errors are rare, so unless you have a need for ECC, you are better served with non-parity (non-ECC) memory.
My Blog => http://adriank.org
Biscuit (03-05-2011)
What in particular is rubbish? Main reason is having ZFS i.e. dataprotection. As I see you can use something from Orcacle (and as I understood currently the developers run away from them) or take something from the community. HAd nexenta running with the four discs without any problems. Interface is not that nice but works.
BtW: Called HP. They will not change anything on the BIOS for the onboard SATA. They say it is for an optical disc and nothing else- Smart guys...
Cheers
Otto
ECC memory really isn't as expensive as it used to be, however you might have to go to a more specialist e-tailer, while memory errors due to faults in hardware are much better than they used to be, memory errors due to the environment are just as high, which on average is 2 faults a year (this is due to solar activity.... yes sun spots!). If you have a machine you leave on all the time, it becomes more likely something critical will be hit. As rebooting the machine normally flushes out the errors, personally I always run ECC memory in any computer able to take it. (I also try to buy machines that can take it (see left)).
Example: Of how silent errors can be introduced, Linux uses any spare memory to cache files "just in case" so its not just programs that crash... If your memory get zapped a bit could be corrupted, so when the cache is used rather than the "good" copy from the disk your get an error introduced, if the cache is used then written to the disk (in for example a new file) that will be corrupted, not knowing the user might then remove the orginal.
If your paranoid about your data. use md5sum to get a hash of its contains. That way when you copy it to a new destination your can verify it.
Last edited by oolon; 03-05-2011 at 02:24 PM.
(\__/) All I wanted in the end was world domination and a whole lot of money to spend. - NMA
(='.*=)
(")_(*)
Yeah I think the average stats are 1 error per 1GB per year caused by cosmic rays - so if you have tons of memory like in a large server it becomes more of a problem.
It looks like you're the only one that's had it running reliably.
For me, running Nexentastor community 3.0.0 it kept crashing, lost data and performance was extremely variable. It turns out the crashing was caused by the web server running on it.
IME, ZFS is great in theory but in practice none of the implementations work properly unless you are willing to spend a very long time testing and implementing workarounds.
Have a look at my comments here http://forums.hexus.net/hexus-hardwa...osing-nas.html
"In a perfect world... spammers would get caught, go to jail, and share a cell with many men who have enlarged their penises, taken Viagra and are looking for a new relationship."
http://www.crucial.com/uk/store/mpar...6974D6A5CA7304
Only about 25% more than non ECC
I've got 2 of these in my server.
"In a perfect world... spammers would get caught, go to jail, and share a cell with many men who have enlarged their penises, taken Viagra and are looking for a new relationship."
The offer seems to have been extended to 31st May.
http://h41112.www4.hp.com/promo/prol...ver/index.html
So I picked myself one up.
Anyone who bought from Crescent received their rebate yet?
Desktop - i7 930, XMS3 6x2GB DDR3, X58A-UD3R (rev2), 2xHD5870 1GB (CrossFireX), Crucial C300 64GB , 2x2TB WD Caviar Green, Corsair 650TX
Notebook - MacBook Pro 13" i5 Early 2011
My flickr
[QUOTE=badass;2074737]It looks like you're the only one that's had it running reliably.
For me, running Nexentastor community 3.0.0 it kept crashing, lost data and performance was extremely variable. It turns out the crashing was caused by the web server running on it.
IME, ZFS is great in theory but in practice none of the implementations work properly unless you are willing to spend a very long time testing and implementing workarounds.
Have a look at my comments here
Hi,
hmm. It tested it for some two months. Worked without issues. Also made link aggregation (have an Intel 2 NIC Card in it). Also looked at Freenas first (as I am using a monowall too it is very familiar). However, ZFS there did not work for me. Win Server 2008 R2 did not work with Raid 5.
In a way it is the never ending story - hardware RAID is expensive and in fact you would need a second controller just in case... So ZFS software raid seems very promising.
Then I came across napp-it and now also build a napp-it all in one with a Supermicro Board, small Xeon and and secong 8xSAS Controller. Also works fine.
What is indeed true: Solaris is very picky in regard to hardware. What did not work in my Microserver was a 5th HDD attached to the onboard SATA in a pool (as it is not AHCI). Now I put the Solaris on the disc attached to the Onbaord SATA and build a pool wit 4x 1TB non 4k discs. Currently doing a Scrub. Ah, and 8 GB ECC also made a big difference in regeard to Performance.
CHeers
Otto
had to remove your url as less than 5 comments yet
I haven't even received the server yet! 5-7 day delivery my rear end! Seriously it's getting pathetic now, still can't get through to them on the phone...
Edit: OK, so I finally got through to them, you just have to be patient and wait a while on the phone. Said they are going to look into it for me as it sounds like it's gone missing or something. Was on the verge of phoning my bank there!
The offer being extended to May is some comfort at least.
There are currently 1 users browsing this thread. (0 members and 1 guests)