I just wonder why we don't learn from past (limits) and remove this limits "forever". E.g. 1 month ago I recieved question of possibility building 10 TB Linux cluster (physics are crazy;-)).
There surely MUST be some way how to do this - I just imagine some file (e.g. defined in LSB) which would define this limits for COMPLETE system (from kernel, filesystems, utils to network daemons). I know there are efforts to things like this but if we'd say (for example) thay that distribution in 2004 won't be marked "LSB compatible" if ANY of programs will use any other limits I think it will create enough preasure on Linux vendors.
I'm afraid it's right. AFAIK even you must have one box per PC AND serial number from the one box must be installed (I dunno how's about some multi licence programs but they probably need more then 20 pcs...)
Hey man - I'm running quite good bussiness selling own developed Linux based NAS which is using SW IDE RAID5 as backup (up to 12 discs!). Till now I've sold about 120 pcs with 23 TB and 800 pcs of discs;-) And I'm selling it to knowledgeable people - SCSI is good for video editing/high end databases but for anything else it's waste of money.
OK, > -What if HD fails? Use RAID and/or store it through NFS to central server where you can backup it. > A RAM failure Use EVERYTIME before deploying node memtest86 - if it pass, RAM is 111% OK.
1000 Mbit - I'd expect that even cheap (Intel Desktop Gigabit - 50$/pcs) should be faster and lower latency than 100 Mbit..
RTT - Round Trip Time - just simply "ping" - time to send packet to node and wait for respondce
Visit - I'd like to;-) But I'm from Czech rep. - where are you from?;-)))
About system reliablity - use "checkpoints" - when one node fails, recover it to last known checkpoint.
About network - why not to use 1000 Mbit? Adapters are VERY cheap today (ok 32 bit PCI only, but nevertheless...) and switch is not SO expensive. And I would expect that RTT should be lower then on 100 Mbit.
About administration - why not use network boot? I'd expect to solve it every problem - you have one image for everybody. Need to upgrade? Just reboot node.
P.S. I've no experiences with any type of clusters;-)
That's quite old - people are doing it for years - only they are using slower technology - something like 50 Bits/sec @ 230 V. Some strange ones are using 110 V.
What exact projects do you mean? I don't know about any such a IMPORTANT project! Look: kernel, bash, Xfree, KDE, GNOME, Mozilla, GCC, GIMP, Samba, NFS, Apache, xinetd, PHP, Perl, Python, Java... should I talk more?
Yes - there are LOTs of abonned projects - but mainly that there is lack of users. And it's even better that such a projects are not lost - if you need it, you can continue where previous ppl stopped.
Re:Three years and counting....
on
Slashdot Turns 5
·
· Score: 1
Yes, I have Windows. But yes, I run Mozilla
Hey man - I'm LIVING from selling and supporting Linux file servers (NAS) for 2 years and I use as my primary and ONLY desktop SW Windows XP;-))) (Mozilla of course;-)
Slashdot Mirror
Just ask - what makes MS money on XBox? Selling more XBoxes? No. Selling GAMES! Will Linux-on-XBox sell more XBox games? No.
The point of my posting is to have limits in only one file for complete system and need to just change it there.
I just wonder why we don't learn from past (limits) and remove this limits "forever". E.g. 1 month ago I recieved question of possibility building 10 TB Linux cluster (physics are crazy ;-)).
;-)
There surely MUST be some way how to do this - I just imagine some file (e.g. defined in LSB) which would define this limits for COMPLETE system (from kernel, filesystems, utils to network daemons). I know there are efforts to things like this but if we'd say (for example) thay that distribution in 2004 won't be marked "LSB compatible" if ANY of programs will use any other limits I think it will create enough preasure on Linux vendors.
Just a crazy idea
I'm afraid it's right. AFAIK even you must have one box per PC AND serial number from the one box must be installed (I dunno how's about some multi licence programs but they probably need more then 20 pcs...)
I couldn't find out nowhere answers to this 2 questions:
;-)
- does ALL SATA adapters + disks supports hotswap?
- does SATA under Linux support hotswap?
And yes, I know www.serialata.org
That makes sense BUT if you'd be unable to play XB1 games on XB2 imagine what "political" power this gives to Sony etc.
Why can't Linux act as XB1 game? Then it'd just skip XB2 key.
3. By that time the hardware would be definitely obsolete, or X-Box 2 would be out with programs signed with a different key
I don't think that M$ could do this - this'd mean you cannot play old games on X-Box 2.
Something common with BrainF*ck language? ;-)
Hey man - I'm running quite good bussiness selling own developed Linux based NAS which is using SW IDE RAID5 as backup (up to 12 discs!). Till now I've sold about 120 pcs with 23 TB and 800 pcs of discs ;-) And I'm selling it to knowledgeable people - SCSI is good for video editing/high end databases but for anything else it's waste of money.
...and switching TV chanell during commercials is theft also?
Writing new IS? I choose PHP ;-) And now maybe I'll check C#.
;-) (in my opinion as a PHP mainly coder ;-)))
But only place where C sucks is in string operations
I'd like to see Bill's face in that moment ;-)
OK,
;-) But I'm from Czech rep. - where are you from? ;-)))
> -What if HD fails?
Use RAID and/or store it through NFS to central server where you can backup it.
> A RAM failure
Use EVERYTIME before deploying node memtest86 - if it pass, RAM is 111% OK.
1000 Mbit - I'd expect that even cheap (Intel Desktop Gigabit - 50$/pcs) should be faster and lower latency than 100 Mbit..
RTT - Round Trip Time - just simply "ping" - time to send packet to node and wait for respondce
Visit - I'd like to
About system reliablity - use "checkpoints" - when one node fails, recover it to last known checkpoint.
;-)
About network - why not to use 1000 Mbit? Adapters are VERY cheap today (ok 32 bit PCI only, but nevertheless...) and switch is not SO expensive. And I would expect that RTT should be lower then on 100 Mbit.
About administration - why not use network boot? I'd expect to solve it every problem - you have one image for everybody. Need to upgrade? Just reboot node.
P.S.
I've no experiences with any type of clusters
Hack /etc/sendfile.cf!
Maybe we should get them together but on some distant island without any connection to rest of world ;-)
Did we really slashdotted MS server? I don't believe it ;-)
That's quite old - people are doing it for years - only they are using slower technology - something like 50 Bits/sec @ 230 V. Some strange ones are using 110 V.
What exact projects do you mean? I don't know about any such a IMPORTANT project! Look: kernel, bash, Xfree, KDE, GNOME, Mozilla, GCC, GIMP, Samba, NFS, Apache, xinetd, PHP, Perl, Python, Java... should I talk more?
Yes - there are LOTs of abonned projects - but mainly that there is lack of users. And it's even better that such a projects are not lost - if you need it, you can continue where previous ppl stopped.
1. Release RedHat 8.0
2. ?
3. Profit!
Yes, I have Windows. But yes, I run Mozilla
;-))) (Mozilla of course ;-)
Hey man - I'm LIVING from selling and supporting Linux file servers (NAS) for 2 years and I use as my primary and ONLY desktop SW Windows XP
"We will beat Linux on clusters. We can't beat them on price, but we have to add value."
You mean THE Linux which is running on several thousands of machines in Google?
It IS PC. Okay - it is crippled PC ;)
hmmm... doom III?