All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
What does "ioping" mean? similar to "dd"?
what is the different between ioping and dd test?
I did an ioping and dd test on my vps, but don't know how to interpret the result. which numbers should I pay special attention to?
what is the benchmark for "excellent", "good", "OK" and "bad" for a vps?
ioping -c 10 /
4096 bytes from / (simfs /dev/simfs): request=1 time=336.8 ms
4096 bytes from / (simfs /dev/simfs): request=2 time=134.5 ms
4096 bytes from / (simfs /dev/simfs): request=3 time=1784.0 ms
4096 bytes from / (simfs /dev/simfs): request=4 time=397.6 ms
4096 bytes from / (simfs /dev/simfs): request=5 time=306.2 ms
4096 bytes from / (simfs /dev/simfs): request=6 time=86.9 ms
4096 bytes from / (simfs /dev/simfs): request=7 time=330.8 ms
4096 bytes from / (simfs /dev/simfs): request=8 time=248.0 ms
4096 bytes from / (simfs /dev/simfs): request=9 time=527.3 ms
4096 bytes from / (simfs /dev/simfs): request=10 time=273.1 ms
--- / (simfs /dev/simfs) ioping statistics ---
10 requests completed in 13431.3 ms, 2 iops, 0.0 mb/s
min/avg/max/mdev = 86.9/442.5/1784.0/462.5 ms
ioping -RL /
--- / (simfs /dev/simfs) ioping statistics ---
1 requests completed in 6149.2 ms, 0 iops, 0.0 mb/s
min/avg/max/mdev = 6149.1/6149.1/6149.1/0.0 ms
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 36.1959 s, 29.7 MB/s
Comments
I'm pretty sure IOping is the random access read writes in milliseconds how long it took to access that data. like a wait time.
that ioping is horrible.
then what is a "reasonable" value for a lowend vps?
ioping measures, amongst other things, latency for i/o requests. Let's take this one for example:
This single disk i/o request took 1.7 seconds. This means that ioping was waiting 1.7 seconds for the operation to complete. Most programs make multiple i/o requests per action, so consider that if every operation took 1.7 seconds, multiplied by the number of requests, would equal the amount of time that the system waits for hardware to finish tasks.
1.7 seconds by itself isn't noticable, but when many operations are concurrently delayed, it will become extremely noticable.
This is one of our heavily-loaded servers, with an LSI 9660 hardware RAID controller:
This is a newer node with fewer people on it, with a Dell PERC H700 controller:
This is a somewhat un-stressed software RAID1 array:
This is from a single hard drive:
This is my laptop:
This is from an extremely overloaded VPS from a company that advertises here on LET:
As you can see, a wide array of responses. It's a good tool to use in considering other data. I will say that your results seem to be a bit poor; does the system feel slow to you?
This is absolutely horrible result. Mind revealing which provider is doing such a great job?
Prometeus, RamNode
ITS OVER 9000
Thanks @Damian.
So the ioping test is for read or write?
Why the single hard drive result seems to be better than RAID1? Is your laptop using an SSD?
I haven't set up anything on this vps yet, so not sure how it will "feel" with a single fresh wordpress site.
In the same sense that the network tool 'ping' is 'round trip' from the source to the destination and back to the source, ioping is write then read.
The RAID1 server mentioned is a shared hosting server, which has customers on it doing various things. My laptop generally only has Firefox and some terminal windows open, so everything gets loaded to ram, and therefore doesn't do much with disk once everything is going.
Say what?...
which provider is doing such a great job?
I think your sarcasm detector is broken, Erawan. <_<'
Hehe.. That's because my problem with English is not my native language
I should write it lke this :
Prometeus and RamNode for great ioping.
16384+0 records in
16384+0 records out
1073741824 bytes (1.1 GB) copied, 36.1959 s, 29.7 MB/s
this is awful, as suppose to be more than 80 MB/s, higher figure is better.
If you have a project that's continually writing 30 megabytes/sec, it would probably be best to rent a dedicated server.
Is there any way to constantly monitor/log the hdd io speed of a vps?
I don't think that would be possible, since to know the i/o speed, you'd need to test it by using 100% of the available i/o bandwidth, which would then impact other services.
Sorry...I didn't make it clear...I mean...how to constantly monitor the actually write speed of ur vps so that u can tell it keeps writing at 30MB/s all the time...any free software/script available?
http://guichaz.free.fr/iotop/ exists, which has a READ/WRITE line at the top. Depending on your virtualization tech, you could install Munin and install the disk plugins; there's plugins for utilization, latency, etc.
You could write a cron using ioping that echoes to a text file. Maybe run the cron every 10-20 minutes.
Maybe the easiest way is using Munin, because some benchmarking script can slow down the node
I apologize for the necroposting, advise Munin plugin for testing disk i/o inside openvz container