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.
Are you getting full disk speed from Oracle free tier?
Hey guys I noticed that some people are not getting the full disk IO speed from their oracle free tier. Below is a yabs example by @sh97, im assuming that was set up with the default IOPS option selected in the boot volume, the VPU slider set to 10. But you can select upto 90 or higher in some cases giving a much better disk speed.
Below is VPU set to 10
Sat Mar 4 17:00:37 IST 2023 ARM compatibility is considered *experimental* Basic System Information: --------------------------------- Uptime : 13 days, 22 hours, 7 minutes Processor : Neoverse-N1 CPU cores : 3 @ ??? MHz AES-NI : ✔ Enabled VM-x/AMD-V : ❌ Disabled RAM : 17.5 GiB Swap : 0.0 KiB Disk : 48.4 GiB Distro : Ubuntu 22.04.1 LTS Kernel : 5.15.0-1027-oracle VM Type : KVM Basic Network Information: --------------------------------- Protocol : IPv4 ISP : Oracle Corporation ASN : AS31898 Oracle Corporation Host : Oracle Cloud Infrastructure (ap-hyderabad-1) Location : Hyderabad, Telangana (TG) Country : India fio Disk Speed Tests (Mixed R/W 50/50): --------------------------------- Block Size | 4k (IOPS) | 64k (IOPS) ------ | --- ---- | ---- ---- Read | 14.48 MB/s (3.6k) | 29.48 MB/s (460) Write | 14.48 MB/s (3.6k) | 30.36 MB/s (474) Total | 28.96 MB/s (7.2k) | 59.85 MB/s (934) | | Block Size | 512k (IOPS) | 1m (IOPS) ------ | --- ---- | ---- ---- Read | 26.35 MB/s (51) | 25.98 MB/s (25) Write | 28.60 MB/s (55) | 28.99 MB/s (28) Total | 54.95 MB/s (106) | 54.98 MB/s (53) Geekbench 5 Benchmark Test: --------------------------------- Test | Value | Single Core | 846 Multi Core | 2468 Full Test | https://browser.geekbench.com/v5/cpu/20805494 YABS completed in 4 min 17 sec
for example here is a YABS with the same 3 core and the VPU setting, set to 90.
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2023-04-23 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Wed Apr 26 12:42:10 UTC 2023
ARM compatibility is considered *experimental*
Basic System Information:
---------------------------------
Uptime : 3 days, 19 hours, 53 minutes
Processor : Neoverse-N1
CPU cores : 3 @ ??? MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ❌ Disabled
RAM : 17.5 GiB
Swap : 0.0 KiB
Disk : 145.3 GiB
Distro : Ubuntu 22.04.2 LTS
Kernel : 5.15.0-1033-oracle
VM Type : KVM
IPv4/IPv6 : ✔ Online / ❌ Offline
IPv4 Network Information:
---------------------------------
ISP : Oracle Corporation
ASN : AS31898 Oracle Corporation
Host : Oracle Cloud Infrastructure (uk-london-1)
Location : London, England (ENG)
Country : United Kingdom
fio Disk Speed Tests (Mixed R/W 50/50):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 63.86 MB/s (15.9k) | 155.83 MB/s (2.4k)
Write | 63.82 MB/s (15.9k) | 160.47 MB/s (2.5k)
Total | 127.68 MB/s (31.9k) | 316.31 MB/s (4.9k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 112.17 MB/s (219) | 109.94 MB/s (107)
Write | 121.77 MB/s (237) | 122.66 MB/s (119)
Total | 233.95 MB/s (456) | 232.60 MB/s (226)
iperf3 Network Speed Tests (IPv4):
---------------------------------
Provider | Location (Link) | Send Speed | Recv Speed | Ping
----- | ----- | ---- | ---- | ----
Clouvider | London, UK (10G) | 3.02 Gbits/sec | 3.01 Gbits/sec | 1.45 ms
Scaleway | Paris, FR (10G) | 3.01 Gbits/sec | 3.00 Gbits/sec | 8.43 ms
NovoServe | North Holland, NL (40G) | 3.02 Gbits/sec | 3.00 Gbits/sec | 6.60 ms
Uztelecom | Tashkent, UZ (10G) | 891 Mbits/sec | 885 Mbits/sec | 105 ms
Clouvider | NYC, NY, US (10G) | 1.21 Gbits/sec | 351 Mbits/sec | 69.7 ms
Clouvider | Dallas, TX, US (10G) | 975 Mbits/sec | busy | 108 ms
Clouvider | Los Angeles, CA, US (10G) | 733 Mbits/sec | 619 Mbits/sec | 127 ms
Geekbench 6 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 1071
Multi Core | 2764
Full Test | https://browser.geekbench.com/v6/cpu/1053867
YABS completed in 9 min 41 sec
.
.
.
.
Increasing Disk Speed When Creating an Instance
When setting up an instance, you can choose the disk speed, for me 90 is the setting that always works (my free tier is in London), ymmv! If i select above 90 sometimes it works and sometimes not.
tick the "specify a custom boot volume size" then change the VPU setting slider for a faster speed disk.
.
.
.
.
For Those With An Already Running Instance/VPS
For those that have already setup their free tier instances don't worry, you can change the speed while the VPS is running, may need a reboot to take effect. To do this follow these steps.
- go to your instance main page
- scroll down and click Boot volume, in Resources
- then click your instance name under the boot volume heading
- click edit
- change the slider to a higher speed then click save.
Comments
Will this affect the cost? Will it remain 100% free or incur additional charges?
It doesn't effect cost
How you guys are using oracle free tier, i was trying from 7-8 months to register and somehow succeeded a week back, and was terminated within 2-3 days... (NO REASON). i was just trying to get free server, it always says out of host capacity.
No it's included in the free tier, but not obvious when you set up an instance.
You want be charged anything on the free tier.
I have tried to sign up several times. Never known such a chaotic thing. Verify button doesn't show in emails, then email "forbidden" due to too many attempts. Second email used, card declined (never happens), start again with second card, temporary payment appears, a neverending spinning thing says the account is being created, eventually close the window, then told no accounts exist and second email is now "forbidden." I sign up with third email and the second card that worked last time is now declined! Finally give up...
The people who design these sites need to be put in prison as they are a danger to the mental health of the population!
The worst bit is this strict signup pathway is the only way to register for their paid services (unless you are large enough to go through sales department). If you fail to hit their invisible free tier requirements, you are also blocked from paid services, some of which are nicely priced
It reminds me of Google with its constant checking and manipulative behaviour. I would rather pay a few Dollars a month and have a good experience.
One of my Oracle idlers just hit the 2 year mark. Fantastic service:
18:33:28 up 731 days, 9:06, 1 user, load average: 0.00, 0.00, 0.00In the OP's YABS, the second has triple the storage space as the first. This will account for 3x of the performance gain as apparently the speed is proportional to size, so the VPU setting might not be making as big a difference as he things.
Nice. Mine are almost a year:
Anyone care to weigh in on disk speed if using paid tier?
Everything above 0 VPU apparently has a cost:
https://www.oracle.com/cloud/storage/pricing/
granted, but whats your point? that it's not worth changing the VPU slider? Might only get double the performance as opposed to triple. It's free double the disk performance.
If this thread was a VPS provider stating please post "i love XXX vps provider " for double disk speed, the thread would be on page 100 already.
Just to prove my point.. on 4k IOPS it's almost quadruple the performance, which is the important factor in most use cases.
Here is 1 core with VPU at 10
same instance with 90 VPU
I think you're looking for an argument where there wasn't really one to be had. I was simply pointing out that multiple people have reported that disk IO is proportional to the size of the block allocated to it.
What relevance has this to anything?
But... if you compare this "VPU at 10" with the one of an allegedly identical setup in the OP, you'll see that this one is half the performance of the OP. So, everything being equal, there's already a 100% variance on the result from the same spec. Doesn't that make you wonder if this metric is actual useful? So are these instances really identical? Was it on the same host? What were the neighbours doing at the time? etc.
Yes and no. If you are below 200 GB (the Always Free limit) it seems you can run whatever the fuck you want. [However UHP is not supported on some VM Shapes)
Above that you pay, a lot, as expected
https://www.oracle.com/cloud/costestimator.html add
Block Volume Storagehow get account oracle cloud ?
https://signup.cloud.oracle.com/
great question
"Registering for Oracle is really difficult."
Thanks, very helpful!
I simply thought people would like to get a faster disk speed at no extra cost, because most people like free stuff and the option is hidden somewhat when setting up an instance.
So you say I'm looking for an argument, then you suggest i might be fabricating information. Seriously what's in it for me? I don't work for Oracle and even if I did I'm not sure they would want me to give away more resources.
It's the exact same instance, i ran YABS 3 times on 10 and 90, they are very consistent. I really don't know what your beef is, first you suggest VPU setting doesn't do all that much, then when I post a YABS with the same instance i'm a potential liar.
All i wanted was to make sure people get the most out of the free offer from Oracle. That seems to have got your goat for some reason.
If you still think I'm a liar, the run the tests for yourself.
I think you might want to step back a bit and read again what I've actually written, as I have never once suggested that you're fabricating information.
What I said is that two VPS instances that are supposed to be identical - same CPU settings, same disk space, same VPU setting - are giving results where one has double the IO throughput of the other. This suggests (as I actually said but you've cut from my reply) that there is something else at play here. I even gave multiple suggestions of things that could be causing that, not one of which was that you were lying.
I don't have any beef with you, nor have I accused you of being a liar. Again, please stop assuming the worst of people and actually read what I wrote.
Whatever dude.
Again, I never once said you were a liar. But in any case, I'm using my two instances, and I'm already happy with their performance, so I won't be fiddling with the setting for the time being.
I messaged them about the experience and I've had a cryptic response. Is it run by the CIA?
Seriously though, not a good omen if even the sign up process for new customers is like this. There's no reason why my cards would be declined.
Thanks for the tip @nocloud
I did the change, before:
After:
@nocloud , thank you for the tips too.
Would there be any negative impact if dialing the VPU all the way to 120?
I hope I can also register successfully.
Not that I know of, but it doesn't always work for me. upto 90 always works and 120 sometimes works. but if you are on the free tier then it all good it won't harm anything to select 120 and if it accepts it all good.