New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
That's right but it has to do a lot with marketing. The bigger number the better. From what I heard some popular hosts optimize nodes specifically to show higher sequential IO throughput speed rates and stupid LEB crowd fall for that.
I do not know about special optimization but caches bigger than 1 GB on the raid controller could be used because the "test" is run usually with 1 GB files.
Challenge: Do the "test" with 1 GB and with 2 and 3 see if there is some difference.
@Heinz agreed, it pains me to see people being willingly sucked in is all and then other hosts come along for an e-penis contest to say that their NV cache is faster than someone else's SSD Cache.
Based on the constant ramnode quality and variation I would suggest they have sequential IO disabled on their cache which is A++ and thumbs up, gg! and all that, sadly not the case for all hosts.
I think a new test standard is needed, will try to work something out.
did I win?
I'm not sure why you invaded this topic to make it RamNode vs Prometeus, but if we need to make it a competition for some reason:
http://serverbear.com/benchmark/2014/01/04/ytvl7PgUVN9pG7wu
Wow, where's that?
@Nick_A agreed sadly that is always the way.
Interestingly look what I get with running Mao's previous test from /dev/shm: (Ram for those that dont know)
Interesting comparison
I don't remember the node, probably a new one at the time. One of our Atlanta users.
Pure SSD or SSD-cached?
Pure
@Maounique simply can't agree how great is RamNode!
@AnthonySmith @Maounique very interested to know a good disk benchmark alternative to dd. is ioping enough?
fio and bonnie++ work well.
In my opinion, he just explained why all these 'dd tests' battles looks silly...
I'm also a happy user of RamNode, Its very fast. Recently I have move to init but i'm not happy with downtime.
Please read above. I was amending the statement of Anthony that nobody can compete in terms of sequential writing.
Further down I say this is a silly competition with no actual good coming to anyone except a few kids that dont know how disks, arrays, SSD and caches work.
1. The providers (those that choose to participate) invest much more than needed for top performance;
2. The customers pay for something they dont need.
Without any tweaks we managed 1.9 on a half empty node. It is not that nobody can compete.
My point stands. Why are you here to make this RamNode vs Prometeus? If it doesn't do any good, kindly leave this thread alone As it is, you come across as just seeking attention.
RamNode vs Prometeus
GETS MY POPCORN
p.s= RamNODER here
Honestly, who cares? I know that there are many providers offering what Ramnode and Prometeus both offer, and they're both stellar providers. I use both of your services.
So STFU, it's not a war here, both of you :P
found this thread and got curious on ramnode's 128mb seattle vps
root@sea:~# uptime
19:13:42 up 199 days, 19:17, 1 user, load average: 0.06, 0.02, 0.00
root@sea:~# dd if=/dev/zero of=iotest bs=64k count=16k conv=fdatasync && rm -rf iotest
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 1.38841 s, 773 MB/s
No longer has the Prometeus vps but uptime was excellent also! Would recommend those 2 without hesitation.
Again I did not reply you, but to Anthony who stated that nobody CAN compete with Ramnode sequential writing. That is wrong, there are providers which CAN compete, not many that would, but I also raised the point that I raised in other places too that this "test" is largely irrelevant.
From what I know this is not your thread (unless ironhide works for you), you have no right to tell me to go, which was also valid even if it was your thread (except offers threads, of course).
Please dont get defensive, it is nothing against you, Prometeus DOES NOT compete in the "dd test", never did, even if our SAN storage can do 700 MB/s after recent upgrades which are still ongoing, that is not the goal, the goal is low io wait and high IOPS count. There are still things to improve there.
This is the happy ramnode customers' thread.
"Trolling is a game ......, albeit one that is played without the consent of most of the players."
Agree with Maonique, we need a new test. I was using H1F and even that was fine for what I was using it for with transfer ranging from 5-35MB/s, so we don't need all that high disk performance TBH. I would prefer lower prices :-)
@Maounique Why are you begging for attention here still? Why are you so uptight? If all you want to do is prove a point about DD, go make a thread about it. Derailing a friendly review thread about another provider is bad taste to say the least.
@Maounique I don't know why you took offense at customers praising a company they like. Be a bigger man, let praise be given where praise is due even if it's not all literally correct down to the dot. RamNode promises performance, delivers it, and customers love it for that plus everything else. I don't think customers expect Nick to operate near-empty nodes that are constantly capable of running 2.0GB/sec dd. I've seen my SSD node slow down to ~500-600MB/sec plenty of times but my page loads tell me actual server load is small enough for performance deviation to be more than acceptable.
When CVPS launched their PureSSD line on E1270s and had better dd speeds than RamNode, I didn't toss out my RamNode for CVPS. In fact, I tossed out the CVPS service and kept my RamNodes because actual page loads were still faster under Nick's management, on a full node. So, RamNode's really not just about dd.
Great performance is a given, so now I look for everything else that might go wrong.
Slow support? Nope. Outstanding no-nonsense support is a given.
Shady business tactics? Nah.
Lazy owner? Hell no. FAQ gets updates, DDoS filtering was added, new routing improvements, new hardware with more space... One look at @NodeStatus on twitter and you know the RamNode team's serious about what they do.
So far the only time I've seen Nick get worked up is when someone wields stupidity as a weapon to bash his brand. I'm gonna make an educated guess that he's mostly too busy taking care of his company and family to go around measuring e-peens the way lots of others do. That's a plus as well.
No one here can compete ramnode in terms of uptime and network using them for 3 month never faced any downtime and network speed is always fast.
month never faced any downtime and network speed is always fast.
So you've been with ramnode for 3 months and your saying nobody else can beat 90 days uptime ?
RamNode++;
RamNode having a nice uptime, so you might need to stay longer
--
Since this thread is about RamNode, I will not make any comment about another provider
I will stay with ramnode for long time.:)
Here we have another proud member of the TL;DR crowd which writes long posts about something they didnt read.
I was not talking about happy customers praising whatever they wish, I was not even talking to Nick_A, even tho he begs for my attention lately, I was replying Anthony. If you care to read what I say, you will see it has nothing to do with happy customers, only with their method to measure something and also about nobody else being able to "compete" in the dd "test". That is all, the rest of your rant is irrelevant.