Only on Experts-Exchange: Sign-up for a free-trial and we'll send you your permanent license!
Here is what you get: 30 Nodes | Unlimited Sensors | No Time Restrictions | Absolutely FREE!
Act now. This offer ends July 14, 2017.
Course Of The Month
6 days, 4 hours left to enrollBecome a Premium Member and unlock a new, free course in leading technologies each month.
Solved
Basic storage buffer and queue question
Posted on 2012-12-30
I have a basic question regarding storage, When data arrive to control unit, it should keep it in queue and send it to necessary lun to write data.
Inside the lun we need to worry about storage pool, raid and disks.
My question is where are the places I need to worry about cache and buffer queue? I know this question is more general, but I'm trying to get the view.
Inside the lun we need to worry about storage pool, raid and disks.
My question is where are the places I need to worry about cache and buffer queue? I know this question is more general, but I'm trying to get the view.
[X]
Welcome to Experts Exchange
Add your voice to the tech community where 5M+ people just like you are talking about what matters.
- Help others & share knowledge
- Earn cash & points
- Learn & ask questions
-
10
6
4- +1
23 Comments
Active 1 day ago
Thank you very much. If want to turn of for specific LUN, can you turn off? Or do we need to do it per storage pool?
queue/buffer on the initiator mean, which initiator? Sorry, you mean by host level?
queue/buffer on the initiator mean, which initiator? Sorry, you mean by host level?
Active today
yes, host level hba initiator.
that would very much depends on the configuration of your SAN, not usually cache is set to all, it's a global.
that would very much depends on the configuration of your SAN, not usually cache is set to all, it's a global.
Active 1 day ago
I'm working on storwize 7000 storage and they were telling me that I can't turn off or on cache on specific LUN?
Active today
Cache, buffer, queue, prefetch, and more exist in multiple places. The more you tune the fewer I/Os have to go to disk, the less frequently you do them; the size of the I/Os are smaller; and your cache has higher utilization..
By far, the HDD is the slowest, least efficient, that can take thousands or millions of times longer to get data when compared to cached data within a host computer.
The HDD has a queue used to reorder instructions. This is not generally programable other than on/off. However cache utilization algorithms; prefetch sizes; the read/write utilization are, depending on make/model. You get the greatest bang for a buck here.
Next step up, the RAID controller (assuming hardware). I/O stripe size and cache on/off is about all that is there, other than RAID, and some deep settings best not messed with. Cache is often binary for a specific LUN.
The HBA/initiator/drivers will have a programmable queue depth, but cache is really only there to prevent interrupts. It isn't something most people are allowed to mess with.
By far, the HDD is the slowest, least efficient, that can take thousands or millions of times longer to get data when compared to cached data within a host computer.
The HDD has a queue used to reorder instructions. This is not generally programable other than on/off. However cache utilization algorithms; prefetch sizes; the read/write utilization are, depending on make/model. You get the greatest bang for a buck here.
Next step up, the RAID controller (assuming hardware). I/O stripe size and cache on/off is about all that is there, other than RAID, and some deep settings best not messed with. Cache is often binary for a specific LUN.
The HBA/initiator/drivers will have a programmable queue depth, but cache is really only there to prevent interrupts. It isn't something most people are allowed to mess with.
Active today
Why would you want to turn the storage cache off, it's backed up onto a local SSD should power fail.
Active 1 day ago
Hi Andyalder,
The reason I wanted to turn of the cache is that I can get better rate on write, when I turn off the cache I got around 680MB/S write, but when i turn of I got 490 MB/S. I wanted to see which option would provide better write. What is the disadvantage of turning off cache?
The reason I wanted to turn of the cache is that I can get better rate on write, when I turn off the cache I got around 680MB/S write, but when i turn of I got 490 MB/S. I wanted to see which option would provide better write. What is the disadvantage of turning off cache?
Active today
The disadvantage of disabling cache is that it slows writes down, which seems to be the opposite of what you are seeing. How are you measuring it and where are you turning it off?
Active 1 day ago
I turn it off on specific LUN and writing data to it and it is increased by closed to 200 MB/S.
I'm using IBM storwize 7000
I'm using IBM storwize 7000
Active today
Increasing by turning off cache is quite rare. Personally, I think these numbers are highly suspect, especially if you are running some benchmark rather than real-world application I/O.
So, specifically, how are you measuring throughput. Not only that, but real-world, are your applications sequential writes? Is this system serving as a data logger?
So, specifically, how are you measuring throughput. Not only that, but real-world, are your applications sequential writes? Is this system serving as a data logger?
Active 1 day ago
I just used DD command to write data which sequential write. One more question, if I'm writing data as 32KB block into raw device and my raid stripe size is 256KB, where does conversion happens? Or it waits for another 8 blocks of data and write all together? Sorry for asking too much questions.
Active today
Exactly what is the dd command you are using?
(But it is moot, as no matter how you use it, there is no guarantee that you will get 100% 32K KB sized CDBs by the time it gets to the disk). In fact, nothing at all even gets sent until a fflush() gets called by the kernel, and that varies depending on some settings.
Also, real-world, this is a crappy benchmark because it doesn't represent a real-world test, unless all your computer does is write zeros in chunks of 32KB to a raw device.
(But it is moot, as no matter how you use it, there is no guarantee that you will get 100% 32K KB sized CDBs by the time it gets to the disk). In fact, nothing at all even gets sent until a fflush() gets called by the kernel, and that varies depending on some settings.
Also, real-world, this is a crappy benchmark because it doesn't represent a real-world test, unless all your computer does is write zeros in chunks of 32KB to a raw device.
Active 1 day ago
I'm using dd command like this. I'm writing in loop.
#!/bin/bash -x
for (( c=1; c<=20; c++ ))
do
`time dd if=/dev/zero of=/opt/testing123 bs=8k count=30000000`
`time rm /opt/testing123`
`sleep 40`
done
#!/bin/bash -x
for (( c=1; c<=20; c++ ))
do
`time dd if=/dev/zero of=/opt/testing123 bs=8k count=30000000`
`time rm /opt/testing123`
`sleep 40`
done
Active today
Other flaws
* You're writing to the filesystem, not hdd. So FS parameters, blocking, journals, all NOT necessarily 8kb sized also gets done. You should be using /dev/sdX if your desire is to test write performance on the hardware.
* You are not writing anything that can be cached by the controller anyway, (except for small amount of housekeeping) that is why turning it off has some benefit. Real-world this just doesn't happen unless your app is a real-time data logger to raw physical contiguous blocks.
* filesystem housekeeping also forces mixed I/O sizes of typically 4KB random reads & writes.
* 8KB I/Os WILL be aggregated by kernel, filesystem, and even device drivers.
* Is /opt a separate partition, or part of the / filesystem?
* You didn't 'nice' the dd, so the test competes with all other system I/O
* Single threaded?? Your test doesn't even let multiple cores share the load. At least run multiple instances in parallel and add a "wait" command to let them finish up. [add & at end of each dd, followed by a wait command to let the n instances finish up, use different output files, set the start= on each instance to insure different parts of the disk]
* dd doesn't sync the output. You have I/O in a queue that has NOT been written to the HDD when 'time' has been run. You should add a "sync" to flush all I/O to disk instead of waiting 40 seconds, and before determining the time if you want to use this script.
* dd isn't even doing contiguous I/O to disk
In other words, garbage in => garbage out. Your flawed benchmark resulted in flawed results that make you think no cache is a good thing.
* You're writing to the filesystem, not hdd. So FS parameters, blocking, journals, all NOT necessarily 8kb sized also gets done. You should be using /dev/sdX if your desire is to test write performance on the hardware.
* You are not writing anything that can be cached by the controller anyway, (except for small amount of housekeeping) that is why turning it off has some benefit. Real-world this just doesn't happen unless your app is a real-time data logger to raw physical contiguous blocks.
* filesystem housekeeping also forces mixed I/O sizes of typically 4KB random reads & writes.
* 8KB I/Os WILL be aggregated by kernel, filesystem, and even device drivers.
* Is /opt a separate partition, or part of the / filesystem?
* You didn't 'nice' the dd, so the test competes with all other system I/O
* Single threaded?? Your test doesn't even let multiple cores share the load. At least run multiple instances in parallel and add a "wait" command to let them finish up. [add & at end of each dd, followed by a wait command to let the n instances finish up, use different output files, set the start= on each instance to insure different parts of the disk]
* dd doesn't sync the output. You have I/O in a queue that has NOT been written to the HDD when 'time' has been run. You should add a "sync" to flush all I/O to disk instead of waiting 40 seconds, and before determining the time if you want to use this script.
* dd isn't even doing contiguous I/O to disk
In other words, garbage in => garbage out. Your flawed benchmark resulted in flawed results that make you think no cache is a good thing.
Active today
I agree, it's a meaningless test, especially using /dev/zero since the storage will de-duplicate it to nothing. Use iometer - http://www.iometer.org/doc/downloads.html and throw some random stuff at it, not just large blocks.
Active 1 day ago
Thank you very much. I'm doing it now using vdbench and will provide you the update.
Active 1 day ago
Thanks a lot guys, almost got whats happening. If I understand correctly, if I`m writing 32KB to raw device, in the storage my raid stripe size is 256KB. How is it going to write in storage? It waits for another 8 blocks (32 X8), and assemble together? And then write? Or it all depends on the storage? what is the general behavior in storage?
Active today
It'll generally put it in the write cache until it can write a full width stripe unless it runs out of cache space or cache is disabled. Since the cache is backed up it can immediately tell the Os write operation complete even though it's not written to the physical platters yet.
Active today
Did you get the same anomalous rsult turning the cache off when you threw random test data at it?
Featured Post
With the new era of mobile computing, smartphones and tablets, wireless communications and cloud services, the USDA sought to take advantage of a mobilized workforce and the blurring lines between personal and corporate computing resources.
Question has a verified solution.
If you are experiencing a similar issue, please ask a related question
In this article, I will show you HOW TO: Suppress Configuration Issues and Warnings Alert displayed in Summary status for ESXi 6.5 after enabling SSH or ESXi Shell.
In this article, I will show you HOW TO: Perform a Physical to Virtual (P2V) Conversion the easy way from a computer backup (image).
Teach the user how to configure vSphere Replication and how to protect and recover VMs
Open vSphere Web Client: Verify vsphere Replication is enabled: Enable vSphere Replication for a virtual machine: Verify replicated VM is created: Recover replica…
How to Install VMware Tools in Red Hat Enterprise Linux 6.4 (RHEL 6.4) Step-by-Step Tutorial
Suggested Courses
Course of the Month6 days, 4 hours left to enroll
707 members asked questions and received personalized solutions in the past 7 days.
Join the community of 500,000 technology professionals and ask your questions.