Amitz
New Member
Last edited by a moderator:
You bastard. Crocodile Dundee had almost faded out of my long-term memory and released whatever brain cells it was holding onto, and now you've swapped it back in and it'll be another 30 years before it expires...
Global, not regional.To me this is the same dream as the regional wifi dream.
And that right there is *exactly* the problem plaguing the VPS market - focusing on quantity over quality. Not that I have any expectations of his foolhardy dream to make it more than a few gasping steps before stumbling into obscurity - but the "most resources at lowest sale price" crap is precisely why fewer people trust VPSes, especially OpenVZ, anymore.the pricepoint of the matter is really what makes something "LowEnd"
Holy tractor trailer hits a trailer loaded with sewer crap which hits a bus full of nuns.I look up to more greenvaluehost not chris.
I want to shake a few things up to change the lowend thing.
future monthly plans include 16gb ram for $7 per month
In fairness, he was discussing his 32MB idea before that thread...in fact, May 1st, right here on vpsboard...Yes, 32MB's is "LowEnd", but the pricepoint of the matter is really what makes something "LowEnd". I assume you made this in 5 minutes after reading the thread on LET.
A ramdisk writeback? Please don't try to host anything. The only reason why memory cache works with hardware RAID cards is because they have a battery backup so if the server crashes, the RAM is kept powered on in the RAID card. If you use standard RAM, it will cause loads of corruption because when the system crashes, the RAM is turned off and therefore cleared. RAM is not persistent data. Let the kernel handle caching to the RAM, it does it automatically with free RAM.2x2tb hdd in sw raid 1 with ramdisk writeback
I just want higher dd resaults as people complain if it's below 250MB/SA ramdisk writeback? Please don't try to host anything. The only reason why memory cache works with hardware RAID cards is because they have a battery backup so if the server crashes, the RAM is kept powered on in the RAID card. If you use standard RAM, it will cause loads of corruption because when the system crashes, the RAM is turned off and therefore cleared. RAM is not persistent data. Let the kernel handle caching to the RAM, it does it automatically with free RAM.
I'm sure people will complain if they have a non functional corrupt file system after a power/forced hard reboot.I just want higher dd resaults as people complain if it's below 250MB/S
so would it be a good idea to only use writeback on /dev?I'm sure people will complain if they have a non functional corrupt file system after a power/forced hard reboot.
should it still be useable?As people have said above, if your using OpenVZ then just leave the OS to use the available ram to cache.
Your going to get performance issues with a 1000 containers no matter what you do.
basic programs like nginx etc... for low traffic websites with cloudflare.Useable?
In what sense?
12.8Mhz per VM:Well let's see.
Ram:
32GB / 32MB is 1024 VM's (That is leaving nothings for host OS nothing for cache)
Disk:
2 * 2TB Disk's Raid 1 Giving you about 100Mb Speed
So each VM would have 0.1Mb I/O Speed!
CPU:
3.2 * 4 = 12.8 GHZ
12.8GHZ / 1000 = 0.0128GHZ per a VM!
Answer:
No probably not as even if you just about have enough RAM for your 1000 VM's, their is going to be absolutely no I/O & CPU available to run something like nginx. Let alone a 1000 copies of an OS.