Tag Info

New answers tagged

1

Besides increasing the number of physical disks, you need to modify Oracle to run more parallel processes. I do this with: NAME CURRENT_MEMORY_VALUE ------------------------------ --------------------- parallel_max_servers 32 parallel_min_servers 8 parallel_servers_target 32 parallel_threads_per_cpu ...


1

You get what you pay for. That server is likely loaded to the brim with other users all making use of the hosting, so there's not much you can do but try another provider or host yourself.


0

You haven't given us enough info for me to tell, but my first hunch would be that you are overusing RAM, and churning swap. with atop, show us the memory listing (press m), including the header section. Don't blame your support. Your kernel and ext4 are very unlikely to be the problem.


0

Your memory use looks high. I can see that you're using quite a bit of swap space at the time you collected that info from top, but it's really the amount of data moving in and out of swap rather than the amount in use that's the issue. I'm guessing this is the case. Is your swap used figure changing from reading to reading? If so, that's almost ...


0

Actually it depends on your application. It doesn't make sense to put a Varnish proxy in front of a Nginx web-server if you serve only static files. Regarding the acceleration of your PHP applications you can already use Nginx (see fastcgi_cache) to apply a caching layer. This could be already a simple solution to gain additional speed. However there ...


0

The two companies can very well limit the upstream/downstream transfer speed with packet shapers. And most likely they do, especially if they're hosting companies. You only have a certain amount of upload bandwidth and a certain amount of download bandwidth. Plus, some hosting companies also meter the actual amount of transferred data, and after a certain ...


1

It depends on how you use nginx. If you are not using any of the caching features of nginx, then you will definitely gain from Varnish. You have to decide what function each will have. This said, if you are looking to do "double-caching", I doubt this is a good idea. Perhaps you could tell us about how your stack is set up...


4

More harddrives generally do increase the theoretical IO performance. By what useful amount differs greatly on the method. In many real world applications (and for smaller numbers of disks) the increase in IO performance is not even close to linear. Scenarios where more disks don't increase the performance are for example those where the limits of the ...


3

Increasing I/O throughput will almost always increase performance. However, there is no magic formula for how many hard drives should be used for a certain number of tables or databases. There are too many factors to consider: data size, schema design, write or read heavy, number of queries, etc. Aside from adding additional hardware it always helps to ...


0

All great answers above. There is another use for the module. When you want to block ads from being displayed in your network, if you simply deny the URLs requested the browser load times suffer (or you get a block message displayed on the page). The solution then is to redirect browser traffic to an internal site serving a blank image (rather than deny ...


0

Solved! This page here contained the answer: http://superuser.com/questions/516030/apache-2-4-on-windows-responds-slowly-hangs-when-serving-some-dynamic-pages I had to use AcceptFilter http none in my conf file. I do not use the firewall they do but it seems to have fixed my issue.


1

sp_who2 can provide some information. But generally the question makes no sense for memory. Memory is MOTLY used as page cache and that is never user specific but always database specific. A data page in the database is not owned by a user. They also get dropped as needed and are kept as long as possible - so if you try to find out why your server uses so ...


1

run sp_who2 active to see all active usage.


0

Hello from the mod_pagespeed team. I think you are right on both counts. js_defer.IRaJtkfeqL.js is potentially render-blocking, although it does get loaded at the bottom of the HTML file. We developed what we think is a good mechanism for ensuring that non-critical resources get loaded after rendering for prioritize_critical_css, but we have not applied ...


2

There's no easy answer since virtualisation is involved. You have a virtual disk sitting on top of a file system on top of a block device presented to a virtual guest which has its own driver presenting a block device to LVM. I don't know for sure if that would necessarily cause such a huge difference, but it may be possible. Beyond that... LVM adds ...


1

In addition to the other two answers here, conntrack (connection tracking) might also be an issue. If you are using Linux and if you are using netfilter (i.e. iptables) your conntrack table might be full. First check if conntrack is enabled. For example: $ /sbin/lsmod | grep conntrack ip_conntrack 51617 1 xt_state $ lsmod | grep -i con ...


1

I recommend you install sysstat package then check the recorded info with sar. sar -n SOCK -s <start_time> -e <end_time> to get the amount of sockets during the benchmark sar -n DEV -s <start_time> -e <end_time> to get network interfaces packets and bandwith sar -d -s <start_time> -e <end_time> to get io stats per ...


1

I would look at file descriptors, possible network/interface saturation, and IO issues. For seeing if the network interface is saturated use iptraf which is a command line tool to view realtime stats. Simply: iptraf For IO issues use iostat iostat 1 that will show the IO usage and load every 1 second. For file descriptor issues use lsof or /proc: ...


2

nginx is merely a messenger vehicle and is unable to help your system to scale if the backend application itself is heavy. 40 concurrent users would be peanuts to any web server, after all. Are you sure that the limit is not in your database / application level? I'm 100% sure the limit is there. Go find it.


1

Bonded interfaces do not grant additional bandwidth to individual network flows. So if you're only running one copy of iperf then you will only be able to use one network interface at a time. If you have two NIC in a lagg then you'll need at least two completely independent copies of iperf running on the computer to see any simultaneous utilization. This ...


2

To your first question - 40mbps sounds about right on tested specification for 2901 ipsec performance with features enabled (ACL + NAT). All the ISR G2 platforms include hardware encryption that is enabled automatically. You wouldn't be getting anywhere near 40mbps without the hardware module kicking in :) If you're looking for a router that can push 100 ...


0

Maybe it's not related to the memory. "Physical" disks on Azure uses what's called Blob storage and that's the reason the performance is not the same as in your local computer even with similar hardware. You can find more information on this link. This seems to be a common problem for the SQL Server community who relies heavily on disk performance. I'll ...


0

In this case, you need to check the basic settings that may not allow the 'Perfmon' to fetch the details on the remote machine. Please take a look at the below suggestion and ensure it is all fine. Turn of AV and Local firewall, to check if the 'Perfmon' collects the value Take a network trace with the help some third party tool and see what's going on ...


0

You don't provide much context around the dmesg timing, but if this is off of a fresh boot that immediately started exhibiting the problem, then your issue is that the switch port is flapping, which could be caused by just about anything (your driver, your hardware, the cable, switch's hardware, or the switch's configuration). Given that you said this is ...



Top 50 recent answers are included