Tag Info

Hot answers tagged

5

I'm almost convinced it your BIOS is limiting your CPUs to 1.20 GHz. In the kernel there are two values: scaling_cur_freq which is the frequency that the cpufreq driver thinks it has requested. /proc/cpuinfo contains this value. It calls the cpufreq_quick_get() funciton. cpuinfo_cur_freq which is the frequency that is actually programmed into the CPU. ...


4

Generally it is not a smart idea. See, databases are written by smart people and use as much memory for caching as possible. If you hit the disc too often, then you have a memory problem - and keeping ANOTHER copy in the same limited memory will not make things better. What someone should do is baseline analysis. What is the bottleneck? Fix it. More memory, ...


4

BIOS changes are the best way to guarantee peak performance but some hardware controlled frequency scaling is inevitable(thermal throttling). Here is a post from Intel about setting CPU Frequency. To a certain extent the CPU frequency can be controlled at the user and OS level. Here is some additional info from Arch Linux about frequency scaling that ...


3

Azure VMs are sized based on their RAM and CPU cores, but all sizes are conspicuously lacking in storage performance, which always becomes the main culprit regardless of how many cores and memory you throw at a VM. Two solutions: For normal usage, just attach as many data disks as possible to your VM and configure them in RAID 0 in the guest system; I/O ...


3

Before trying to tune Apache make sure you have a proper monitoring in place so you can see how many requests Apache serves and how the system resources correlate to each other. An easy to use solution for this is munin. Next, prepare load tests with which you can put your system under load to see what effect the different tuning options have. There are ...


3

A TTFB between 400 - 600 ms may be normal for non-optimized servers / code, but it certainly is not the best you can get. You can optimize your server with Nginx, Varnish, HHVM or Redis (or multiple) to make both static content and dynamic content load faster. But don't look only at the TTFB, there are other numbers that matter.


3

Fairly new to Azure, but here's my offering: The more Azure disks you span the better your performance will be, this is true. I'm not sure about the equation though. Either way, I think you're barking up the wrong tree... Azure imposes artificial limits on metrics such as IOPS and Max reads/writes per second, so the queue length is only one of a few ...


2

You're focusing a bit much on a single metric! Diagnosing a bottleneck to a single component rarely ends with one counter giving a full explanation. There are quite a few great guides for using perfmon to diagnose performance problems on SQL Server. And unfortunately your Admin could be right, the counter you choose does indeed depend on the underlying ...


2

Yes. If in doubt, try checking the time to first byte of popular websites in your market, or take a look at this article : https://moz.com/blog/improving-search-rank-by-optimizing-your-time-to-first-byte If your ttfb is between 400-600ms, you have nothing to worry about on that score. Although you may possibly be able to shave off a few ms, if you are ...


2

KeepAlive normally makes a lot of sense, requires more memory but lowers the number of connections, CPU usage and connections overhead. MaxClients and other stuff must be tuned to your situation instead. Normally you'll want Apache to have a number of idle instances big enough to serve all of your users without having it to spawn new child processes all the ...


2

As you described it, that would be the equivalent. 8 cores = 8 vCPU's on HyperV. I would ask/state the follow up question/statements: 1) make sure your physical host has at least 8 cores. 2) does your guest really need 8 vCPU's?


2

When you're using "localhost" notation in fact MySQL uses UNIX domain sockets. They're more performant, comparing to TCP/IP, and for high-load it can make noticeable difference. But UNIX domain is local to the host by its nature, it's not networking, so this implies you can't use remote MySQL server with it.


2

I'd recommend option 2 for just RAID 1 only. It's not really worth it to complicate your setup and double your chance of failure. Probably your best bet for improved performance would be to improve the rest of your hardware (though not an option you say). If your RAID0 fails, it will almost certainly just fail and not work at all. I can't really say what ...


1

The better start can be the set of following rules: disable logging and accept_mutex enable sendfile set sendfile_max_chunk Configuration: events { accept_mutex off; } access_log off; sendfile on; sendfile_max_chunk 512k; New Nginx (1.7.11 or newer) feature thread pool can be really helpful in your case: location / { root /home; aio ...


1

It's very hard to make anything out of your question. I decided that your question is Will a remote database at my provider that I can access via mysql.example.com be slower then a local database that I run on my machine/VPS and that I can access via localhost. The answer is: It depends on many things, among them The specs of your own server/VPS ...


1

It's not entirely clear to me what you're asking, but I get the impression you're talking about a web server like Apache or nginx running on the same leased server as your MySql database in a hosting provider's data center. You're worried about the connection from the web server to the database, and the only difference between the two choices is what name ...


1

They use a dedicated database server because it's easier to manage. Having dedicated database servers and web servers makes it easier to update, easier to backup, etc. If you have everything for a single customer on a single server you have to run updates and backups on every single instance. By separating them you can have as much customers on a given ...


1

A database server is a database server; regardless of where it's sitting. It's fairly easy to just repoint a line of code to a new database server. As with any technology related investments, you have to consider the following: latency; how much tolerance can your app or your company manage? performance; will the replacement solution provide enough ...


1

Write cache: disabled. Do you want the write cache disabled? If not, execute hdparm -W1 /dev/sdb and verify it was changed with hdparm -W /dev/sdb. Rerun your test. If it looks better add write_cache = on to /etc/hdparm.conf.


1

You shouldn't use your c drive for applications: http://blogs.msdn.com/b/igorpag/archive/2014/10/23/azure-storage-secrets-and-linux-i-o-optimizations.aspx The C drive is optimized for boot time, not for high IO. As each disk is IOPS-throttled (normally 500/disk), you can either use many of them (up to 8 on A3) or us a machine from the d-series with SSD ...


1

I have spent the last year (off and on through 2014-2015) testing several parallel CentOS 6.6 RAID 1 (mirrored) configurations using 2 LSI 9300 HBA verses 2 LSI 9361-8i RAID controllers with systems built on the following: 2U Supermicro CSE-826BAC4-R920LPB chassis, a ASUS Z9PE-D16 motherboard, 2 Intel Xeon E5-2687W v2 Eight-Core 3.4 GHz Processors, mirrored ...



Only top voted, non community-wiki answers of a minimum length are eligible