Gamin/gam_server eating 70%+ of my CPU

Searched for this, no solutions on the forum that I could find. Noticed gam_server starts to eat up most of my CPU when I've had the PC on for a while. Not entirely sure what it does, if I can uninstall/disable it, but it's a pain in the ass when simple web browsing gets dragged to a near halt twice a day or so.
Any advice?

In my experience gamin works a lot better (with thunar) than fam does. For me both in ubuntu (which I used before moving to Arch) and in Arch sometimes fam starts to use close to 100% of one of the CPUs for no special reason. fam does provide more functionality than gamin does though, I was told that fam is a lot more efficient if you use nfs and want to keep the file list updated, but as I don't use nfs gamin works much better for me.
File alteration monitoring (with thunar) seem to be a little broken with fam if you login graphically with two (or more) separate users, only the last user to login will get file changes notifications, while with gamin it will work ok for everyone. Also when thunar hangs and you kill it (usually I use "killall Thunar") if you are using fam you need to restart the daemon, with gamin you just need to open thunar again.
Keep in mind that some programs may behave a little awkwardly, one example is medit, latest packages it will not monitor file changes properly if you use gamin, you will need to recompile it from abs changing the PKGBUILD (remove the build option --with-fam, will file a bug or something later when I ask around the best way to suggest that change).

Similar Messages

Plugin container is eating 90 percent of cpu

plugin-container is eating 90 percent of cpu when we usually look videos in youtube. Earlier this was not happening. It started in last two days. Kindly request you to solve this problem
== This happened ==
Not sure how often
== running flash movies

My computer nearly stops since updating Firefox (and gaining this Plugin-container.exe abomination). Come on Firefox, I have changed from IE because it is rubbish don't make me think the same about you.

Flash player hangs in Firefox. Eats up 50% of CPU

I'm working with flash application. It works in other browsers except for Firefox. In Firefox, after couple of minutes, the application hangs and Flash player eats up 50% of CPU. Firefox becomes unresponsive and stays like that. Neither Firefox nor Flash player crashes. Flash just hangs forever. I had to go to task manager to kill Flash player process.
This doesn't happen when I install a Debug version of Flash player or one of the older versions of Flash players. Is there a known issue with new Flash player?
I'm using:
Flash Player 11.5.502.146
Firefox 18.0.1
Windows 7 32 bit.

I'm sorry you're going through this. I suspect you're running into a conflict with our Protected Mode which was introduced in Flash Player 11.4. Please try the troubleshooting steps outlined in this FAQ:
How do I troubleshoot Flash Player's protected mode for Firefox?
In particular, I'd recommend temporarily disabling protected mode (for troubleshooting purposes only) to see if that addresses the issue. If this does fix the problem, please open a new bug report on this issue over at bugbase.adobe.com. The following FAQ will walk you through the information we'll need and instructions for generating a crash log (if applicable):
How do I report a Flash Player crash to Adobe?
Once the bug has been added, please post back with the bug URL or bug number so that others effected can add their comments and votes.

Compressor eating up too much cpu

hi guys,
everytime i use compressor on my macbook pro i7 17 inch, my final cut gets very choppy as well as other apps running. I think its because compressor is eating up too much of my CPU power. is there a way that I can still compress stuff in the background while setting it to be using less CPU?
Thank you
Santo

jackkk wrote:
no I use a reference fcp movie to render. my media files are in external drive connected through esata. im encoding to mpeg2. my source media is aic
A reference movie refers to every frame of every clip in the project. If any of those clips are in use in another open application, you're going to see a performance hit of some kind. Try exporting your timeline as a self-contained movie first. There are reasons why reference movies are efficient, this may not be one of them.
bogiesan

[SOLVED] X eating 60% of the CPU

Hi everyone,
I finished installing Archlinux with the AUR RT kernel, everything works (damn) fine, but X. No matter what I do, or don't, it eats around 60% of the first CPU. I saw some post about it, but the answers to them doesn't change a thing on my setup.
Any help on this one ?
Regards,
Daniel
Last edited by DanielD (2013-10-08 20:39:58)

Please remember to mark the thread as solved https://bbs.archlinux.org/viewtopic.php?id=130309

Firefox is eating up between 90% CPU usage to %230 CPU usage, while on a text-only page. I've tried many of the fixits. They didn't work. What can I do?

I've cleared cookies multiple times
I've emptied the cache
I cleared my download history, and a majority of my browsing history
I've tried having hardware acceleration for both flash and firefox on and off
I turned off all my apps and plugins and when that did nothing I turned back on the ones I need. I never had a lot of apps to begin with. only a max of 5. I had quite a few plugins, but i disabled all but quicktime, flash, and silverlight. I have those three set on "ask to activate" so none are "always active" anymore.
I also tried a couple things from the FAQ and forums here about changing a few settings in the "about:config" setup.
A few of the above things initially worked, but after a day or two it'd be back to sucking up my CPU and memory. I would then try the same thing again, and this second time there would be no improvement.
The real memory will fluctuate between 200 MB and 1.7 GB, and has done so for quite a while.
Even after I've left streaming sites, and sites with lots of images, and I'm not downloading anything, the memory and the CPU will still do that. I only have a handful of tabs open at a timeusually, and all of them are mostly text. I'll quit and reopen firefox, which will initially bring the memory and CPU usage down, but it inevitably climbs up quite quickly, regardless of my tab usage. It'll be the same if I'm playing a flash game or scrolling through blogs.
I never had any of these problems with Firefox 23, and my internet usage habits haven't changed at all. I had some of this slowness a few times with version 24, but nowhere near as bad as this, or as constant. Almost everyday I end up googling for other ways to speed things up. Everything I do never slowed down Firefox on earlier versions.
I've tried all my usual tasks on Safari which will never climb above 30-40 MB of memory, even on netflix streaming, or flash games, and the CPU will average around 8 and 11%.

i'm also having major CPU problems w/ FF 29.0.1. when i restart the app it calms down again but after loading tabs it steadily climbs up to 100% after not too long. i have checked out the 2 CPU links/articles above and do all the basic troubleshooting and tips, but with all the FF tools & add-ons isn't there some sort of CPU monitor that has the ability to display which extensions/scripts are hogging the CPU? Activity Monitor on Mac just shows the total used by the app. it's impossible in my case to efficiently disable dozens of add-ons & scripts, restart, and test each one by one b/c it can take several hrs before the CPU starts to spike high again. i have no time for this.
thanks

Launchd eating 100% of a CPU, causes machine to permanently hang

Gentleones,
I'm attempting to create and schedule a script that will rotate the apache2 logs in a nice fashion like all the other logs on the machine. (Why Apache can't do logging like everybody else, I don't know, but I guess that's a little off-topic.
So I naively figured I'd just write a script to roll the logs and use launchd to schedule its execution. Far below, you'll find the plist that I created as well as the rolllogs.sh script.
When I manually execute the script (using sudo so it can play with the logfiles and whatnot), it runs fine. Gzip fires up occasionally and awstats updates like it should. The problem comes when I try to use launchctl to get this script going. When I use
launchctl load net.eccles.rolllogs.plist
CPU usage goes immediately to 100% on one CPU (my signature list of "Model" is way out of date) and I can no longer get anything else to run. I can't open a new ssh into the machine, I can't use top unless it's previously running in another ssh, and I never get a command prompt out of the session that is running launchctl. It goes off into la-la land (adjacent to one of the "stans" near the Black Sea, from what I hear).
If I reboot the hard way (5 second power button, because nothing else works), then the machine is back up and running. But when I try to log in again on the machine with the GUI, the machine immediately goes off into la-la land with 100% CPU usage again. The only way around things is to ssh into the machine and remove the plist from /Library/LaunchAgents where I put it in the first place. Then launchd can't find it and it doesn't try to leave the country.
So... I figured it's my script that is causing the problems, right? Well, to sort this one out, I created a simple script, shown way below, called simple.sh which does nothing more than create a line in the system.log file. It, too, works just fine if I call it manually. But when I try to get launchd to run it, same problem as before and nothing appears in the system.log file.
My guess? There's something hokey about my plist file that is really screwing up launchd, but I can't find it.
Anybody got any ideas? Anybody had any success with their own launchd scripts like this one? (Note that I'm trying to get this to run on Sundays at 0315, so I think I have the plist entry correct for that.)
Thanks,
Bill
P.S. I discovered while making this post that square brackets are important somehow to this forum software, but you can escape them with a backslash.)
P.P.S There's lots of whitespace in the plist below, but it gets eaten by aforementioned forum software.
shr-g5:~ admin$ more net.eccles.rolllogs.plist
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple Computer//DTD PLIST 1.0//EN"
"http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>Label</key>
<string>net.eccles.rolllogs</string>
<key>ProgramArguments</key>
<array>
<string>/usr/local/bin/rolllogs.sh</string>
</array>
<key>StartCalendarInterval</key>
<dict>
<key>Day</key>
<integer>0</integer>
<key>Hour</key>
<integer>3</integer>
<key>Minute</key>
<integer>15</integer>
</dict>
</dict>
</plist>
shr-g5:~ admin$ more /usr/local/bin/rolllogs.sh
#!/bin/bash
# Update statistics for AWStats...
cd /Library/WebServer/awstats/wwwroot/cgi-bin
for i in `ls *.conf | sed -e 's/awstats.//' -e 's/.conf//'` ; do
/Library/WebServer/awstats/wwwroot/cgi-bin/awstats.pl -update -config="$i"
done
cd /var/log/apache2/
for i in `ls *_log` ; do
echo "$i";
if \[ -f "${i}" \]; then
echo -n " ${i}"
if \[ -x /usr/bin/gzip \]; then gzext=".gz"; else gzext=""; fi
if \[ -f "${i}.6${gzext}" \]; then mv -f "${i}.6${gzext}" "${i}.7${gzext}"; fi
if \[ -f "${i}.5${gzext}" \]; then mv -f "${i}.5${gzext}" "${i}.6${gzext}"; fi
if \[ -f "${i}.4${gzext}" \]; then mv -f "${i}.4${gzext}" "${i}.5${gzext}"; fi
if \[ -f "${i}.3${gzext}" \]; then mv -f "${i}.3${gzext}" "${i}.4${gzext}"; fi
if \[ -f "${i}.2${gzext}" \]; then mv -f "${i}.2${gzext}" "${i}.3${gzext}"; fi
if \[ -f "${i}.1${gzext}" \]; then mv -f "${i}.1${gzext}" "${i}.2${gzext}"; fi
if \[ -f "${i}.0" \]; then mv -f "${i}.0" "${i}.1" && if \[ -x /usr/bin/gzip \]; then gzip -9 "${i}.1"; fi; fi
if \[ -f "${i}" \]; then mv -f "${i}" "${i}.0"; fi
fi
apachectl graceful
done
shr-g5:~ admin$ more /usr/local/bin/simple.sh
#!/bin/bash
logger TestMessage

Here's the CORRECT plist. There's one subtle difference, "Day" vs "Weekday." It's enough to cause launchctl to go out to lunch as described above.
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE plist PUBLIC "-//Apple//DTD PLIST 1.0//EN" "http://www.apple.com/DTDs/PropertyList-1.0.dtd">
<plist version="1.0">
<dict>
<key>Label</key>
<string>net.eccles.rolllogs</string>
<key>ProgramArguments</key>
<array>
<string>/usr/local/bin/rolllogs.sh</string>
</array>
<key>RunAtLoad</key>
<true/>
<key>StartCalendarInterval</key>
<dict>
<key>Hour</key>
<integer>3</integer>
<key>Minute</key>
<integer>15</integer>
<key>Weekday</key>
<integer>0</integer>
</dict>
</dict>
</plist>

How can i improve jsf application? It eat a lot of CPU.

I use only request beans, state save method is server.
How can I improve it, what is with state manager, save state tree.?

Profile your application and find the CPU hotspots. Then you can figure out what needs improvement.

Safari eating up 75% of CPU

Hello,
I would like to know if anyone has any suggestions on what is wrong with my browser. The past couple weeks Safari has slowed drastically and I can't really figure out why. I have emptied the cache, and deleted all of the favorite icons a few times.
When running activity viewer, it uses up about 75% of the CPU when loading pages. My mouse gets slow and unresponsive, along with the pages taking much much longer than normal to load. The only other programs running are typically iTunes, Mail, and Adium (instant messaging). Those programs run fine except iTunes may bog down some if I'm waiting for a page to load in Safari. Anyway, hopefully someone has some suggestions for me. Thanks a lot,
- Shawn
PB G4 15 AL Mac OS X (10.4.7) 1.5 GHz, 512 MB DDR SDRAM, 20GB iPod, Airport

Shawn --
Try these and report back, OK?
Do these & report back, OK ?
1. Empty your cache & History settings via the Safari menu
2. Trash your Icons Folder:
HD>Users>Home>YourName>Library>Safari -->Icons
Restart Safari and a new folder will be automatically regenerated.
(This folder saves favicons of all the sites you visit. Needs cleaning periodically -- or you can change that file's preferences to "Read Only"
rather than "Read & Write." This will keep it from growing in size.)
3. Launch Font Book and check all fonts for duplicates. A duplicated font will have a bullet beside the font family and you can resolve duplicates by selecting the font and at the menu bar, go to Edit > Resolve Duplicates.
Do this for each font with a bullet next to it.
If you're still having problems:
4. Shut down your computer . . . unplug the computer, modem, and router, if you have one. Let everything sit for 2-3 minutes. Plug everything back in, start up your modem & router, then restart your computer.
5. Do you have internet apps SpeedDownload, iGetter, PithHelmet, Acid Search, Safari Enhancer ? Or Norton or Virex? Sometimes non-apple internet apps need updating or uninstalling. If you do, let us know which ones and which version, please.
Hope this helps

Vanguard.exe eating 99% of my CPU

Hi,
I've had a couple of frustrating days. Performance on my PC (running 2000 Pro) is non-existent. When I look at the processes page of the task manager, I have one of two scenarios:
Vanguard.exe is taking 99% of the CPU
-or-
Vanguard.exe is taking 49% of the CPU and Startup.exe is taking 40-46%.
It is my understanding that both of these files are associated with Linksys.
Has anybody encountered this? If so, can you explain what is going on and how I might resolve the problem(s)?
Thanks!
Rickpgh

Hi,
I've had a couple of frustrating days. Performance on my PC (running 2000 Pro) is non-existent. When I look at the processes page of the task manager, I have one of two scenarios:
Vanguard.exe is taking 99% of the CPU
-or-
Vanguard.exe is taking 49% of the CPU and Startup.exe is taking 40-46%.
It is my understanding that both of these files are associated with Linksys.
Has anybody encountered this? If so, can you explain what is going on and how I might resolve the problem(s)?
Thanks!
Rickpgh
=============
Can you post what Linksys devices?
Chris

Swing application eats CPU on resume after standby

My application contains a mix of regular Swing components and about 200 (small) custom components (JPanels) that draw themselves. The problem is that if put my laptop into Standby, then restart it, the application eats CPU. It goes from ~4% pre standby to ~48% after standby. Totally repeatable. There is only the main thread, the AWT thread and a couple timers firing: one to get data from a server and one to redraw the custom controls to make the control blink (calls repaint() to cue up a call to paintComponent()). I put timing code in the two timer driven threads and both are running fine. NOTE: each of the custom components is just a rectangle with a short text string in it. Think checkerboard with text. The background color is all that is being changed. It's basically a grid of alarm/status "lights".
I did some searching and only found a reference to a drawing bug that was fixed back in 1.4.x that had to do with WinXP and dual-headed computers.
Note that another simple Swing application I wrote (no custom controls), doesn't appear to behave this way, so I suspect it has to do with the custom drawing stuff (rolling my own paintComponent() method).
Any thoughts on how to figure out what is eating the CPU? I'm using NetBeans 6.7 and JDK 1.6_14 under Windows XP.
Edited by: garrysimmons on Sep 15, 2009 1:52 PM
Edited by: garrysimmons on Sep 15, 2009 2:57 PM
Edited by: garrysimmons on Sep 15, 2009 3:02 PM

It doesn't matter if it's a Swing timer or a Util timer. It behaves the same either way. That was one of the first things I tried.
I ended up refactoring the code to go from a grid of individual controls (think checker board where each square is a control that draws itself), to a single "board" control that draws a grid. That seems to have fixed the issue. I can go in/out of Standby and the app doesn't eat a ton of CPU any more. The real problem may still be there, just too small of an effect to notice.
I don't know why the original design would work fine before a Standby, but not after. Having ~200 little controls doing repaints twice a second must confuse something in the JVM/Windows versus having 4 large controls doing the same amount of drawing.
It's working, so time to move on.

Textedit won't open. Hangs eating 100% of CPU, works on other User account

Hey everyone,
This is driving me mad. Text edit will not open, if I click on it, it doesnt do anything, just a spinning beachball and the textedit process eats 100% of the cpu. I have to force quit the application. I have tried trashing the textedit plists, zapping pram, even unpacking a textedit app from the Snow leopard install disk using pacisifist. And of course, I have repaired permissions and disks. NOTHING is working. I made a test user account, and text edit works perfectly there. I am at my wits end, any ideas?
Thanks!

HA! Thanks for the response, but I finally figured it out last night. You'll never guess what it was. I was just fed up and let the computer chug away at 100% for about five minutes. FINALLY the text pad came up with 1,400 page document that was just total gibberish...guess it had it from a cache, but god knows from what. Anyways, just closed that and all was good. Thanks for your help though!

Kernel_task eats CPU when MIDI system wakes up

I have a MacBook Pro and a Focusrite Saffire (connected via firewire).
Noticed that the latest 10.5's are eating a ton of cpu when running Logic (or Ableton Live, my other DAW).
Digging around, using Activity Monitor I see:
when I plug in my Saffire, kernel_task jumps from 2% cpu to about 7%.
when I start Logic/Live, it jumps to 45-50% cpu and stays there (regardless of whether or not sound is being input or output).
Digging a little deeper, with no DAW running, I fire up Audio/MIDI Setup and switch the default output to the Saffire... kernel_task steady at 7%. Click on the MIDI Devices tab and kernel_task jumps up into the forties. Unplugging the Saffire, the cpu load goes down. Replugging, it does not go back up until I click "Rescan MIDI".
I have no MIDI devices connected to the Saffire (nor do I have any midi cables with which to try that experiment).
Using dtrace, I could probably dig more but I'm not on Apple's payroll and I'm too old to hack for free. I'd rather make music.
Fingers crossed that this can be fixed cause half a core is better used in a soft synth or something.

Seems like the problem persists when playing a 720p video on the external display set to 720p. To sum it up
Video Resolution
Main Display
External (1280x720)
External (1920x1080)
480p
ok
fluctuating kernel_task cpu usage (~5 to ~150 cpu usage)
high kernel_task cpu usage
720p
ok
high kernel_task cpu usage
high kernel_task cpu usage
1080p
ok
high kernel_task cpu usage
high kernel_task cpu usage

High CPU & Memory issues - Excessive and unnecessary CPU usage in Firefox for Windows (see also questions1003966 & 998289)

I am using the latest version of Firefox, 32.0.1, on an HP Pavilion dv7-4141sa laptop computer, running Windows 8.1 Pro 64-bit. There should be more than enough power here to run a browser without any problem, yet the computer is regularly being overloaded by the browser's resource usage.
I am currently measuring typical average CPU usage of 45% - with the browser minimised. That is to say, Firefox is not doing any browser jobs - it is not drawing on the screen - but it is using the most CPU of any process currently running. Why?
I am an ex-software engineer, and my understanding of current application development technology is limited, but I think I can say with some confidence that many techniques and software architectures exist that enable the suspension of program activity, or at least its reduction to a very low bare minimum, when the program is not actually carrying out its design function, which is in this case to draw pictures of website data on a screen.
There is a very serious problem here, one which means that I cannot run Firefox all the time, but have to kill it when I am not actively browsing, otherwise other applications cannot run to their full capacity. This is a great pity, because I have committed a lot of effort to making Firefox my primary web tool. It manages all my many bookmarks, contains my secure password manager (LastPass), manages my web security (via NoScript, AdBlock, and FlashBlock), just for a start.
Before any well-meaning moderators or other commenters start down the usual trail of "have you done this, tried that, disabled the other", I should point out that I have followed all the advice I can find, with no positive outcome. This computer is running a fresh installation of Windows, and a fresh installation of Firefox. At the moment I am running a new test Firefox profile.
The latest version of Adobe Flash, 14.0.0.179, is installed but disabled. When it is enabled and in use, Firefox's CPU usage seems high over longer periods, but overall there is little difference. It doesn't look as though the problem can be laid at Flash's door. A few other add-ins are installed, mainly those mentioned above, and Firefox is skinned with LavaFox v2-Green, but I have previously started from a clean installation, which has made little difference after an initial period during which things seemed to be better. Turning off or disabling various add-ons has made no discernible difference.
Just now I turned Flash on, and after a long period with Firefox minimised, Resource Monitor reported Firefox at 8% average CPU and FlashPlayerPlugin at 0.32%. While running a small video, RM reports FlashPlayerPlugin at about 45% and Firefox at 8%. After the video, Flash went back down to around 0.4%, and after a bit more browsing Firefox went to around 17% minimised. There seems little rhyme or reason here, except that Flash does not seem to be the culprit, although Firefox does seem to consume less when fewer tabs are actually active, as you might expect. I just refreshed more than a dozen tabs, and now Firefox is at 30%, but still not back up to the higher figures. Weird.
The CPU usage does vary a lot around the average, but 30% is still a lot for a program that shouldn't be doing anything above the minimum status maintenance, and still puts it at the top of my Top Process Monitor list. (Right now it's gone back up to 45%, without my using it in any way).
I have Telemetry turned on, so presumably Mozilla is recording some of the relevant data.
This is a very serious fault, which is more or less reproducible and which a lot of people seem to be experiencing. Probably more people have it than realise, because most users will not know anything other than that their computer slows down for some non-obvious reason; they may just think it's something to do with the computer itself, not knowing how to use the Resource Monitor and interpret the results.
Oddly, the problem does not seem to occur when running Firefox under Linux Mint 17, where the response is snappy and the CPU usage ostensibly low; I am going to look at this in more detail shortly.
I am not posting here for advice, but for information and as a form of bug report which hopefully other users will see. It will hopefully spare them a lot of effort and time in trying a lot of recommended "solutions" which will not work. The problem is inherent in the software, and can be solved only by Mozilla.
I am willing to supply any diagnostic data requested; as I said above, I have also turned on "Telemetry". However, I will not be trying out any more reinstallation/new-profile/add-on disable-enable-delete cycles unless I am asked to for very specific reasons. Please do not suggest courses of action which have already been mentioned elsewhere on this forum. I have tried them all.
The situation is very disappointing, but I have confidence that the Firefox developers will pin down the fault eventually and engineer a satisfactory solution that makes it viable to leave Firefox open all the time. Even with that restriction, it's still far better than the alternatives.

I share jmward's frustration. This is not a bug report so much as a request for a badly needed feature, ie the ability to have Firefox be able to tell us which tabs on which windows are eating time so we can target them. There are some very poorly written web pages out there which just plain loop while making no updates or inconsequential updates on the page.
This is not a OS problem; jmward runs Windows and I run Linux but we have the same issue. It is not a question of Firefox plugins or extensions, or even viruses; I have a new, clean Linux install but have the same problem on both my regular Firefox AND on a FRESH install of the latest Firefox (35.0.1) which has *NO* non-default plugins or extensions.
Here is an example of a web page which will eat 100% of one CPU's time:
http://patch.com/massachusetts/arlington/arlington-woman-allegedly-attacks-neighbor-snow-blower-0
I will attach 3 images to back up my claim. All of the images show my dark blue system monitor (gkrellm) superimposed on the web page. The top graph shows CPU time. I have 4 CPUs in my machine, so 25% means the web page is eating 100% of one CPU equivalent.
The first image shows the load of the offending page with elevated CPU usage along with the download net activity on eth0, but the CPU activity does not stop when the download activity is over.
The second images show the steady state 100% CPU equivalent of this page just sitting there apparently doing nothing. It IS, however, doing at least two different things. I know from my regular Firefox where I have "Remove It Permanently" installed that if I remove the the little weather indicator in the upper right of the offending page that CPU utilization drops to about 40% of one CPU equivalent. If I use RIP to remove everything else right down to a bare screen that 40% utilization stays until I delete the tab.
The third image shows the CPU utilization drop to zero when I delete the tab holding the offending page.
Now I will add a note here: If the offending page is NOT the top tab in a window then it stops eating time. The offender will also stop eating time if I minimize the window. However if I have another window up but leave the offender up on top in its window it continues to eat time even tho I am looking at another window. The offender will continue to eat time even if its window is completely covered by another window, and it will also eat time if I switch to another desktop.
I will often have half a dozen windows open, and there are many pages out there which will suck endless CPU time. Unless I want to open up a new tab on each window before I leave it there is likely to be some tab on top eating time uselessly. Get a couple of those going and Firefox becomes sluggish and can even stop responding for many seconds.
I guess I not only want to have Firefox tell us which tabs are eating time, but I would also like it if Firefox would stop giving CPU time to any tab which is not visible because it is covered by some other window and/or it is on a desktop which is currently not showing. If you can stop a tab's CPU usage by reason of tab-not-showing then you should also be able to stop its CPU usage when the window is covered or on the not-top desktop. Indeed is there any reason to not schedule absolute CPU utilization priority to the visible tab on the window with focus and give time to the other tabs only when the visible tab isn't using time?
My final thought on this is that I have had Firefox SO tied up sometimes that I swear their are web pages which are eating time even tho they are not the top tab. I regret that I do not have an example to show you. But Firefox grinds to a complete halt on a regular basis, and that is not acceptable. I am a programmer with long experience in multi-threaded programs so I *know* there is no reason you cannot do better. If there were another browser around which had the commitment to privacy which Firefox has I would probably switch out of shear annoyance. :-(

Dbwr consuming high CPU after enabling DirectIO

Hi,
DBWR is consuming high CPU. After enabling DirectIO on Solaris SPARC 10, dbwr is eating away almost 1 CPU on a v440 machine i.e. 19% throughout the day. Neither of "buffer busy waits" or "write complete waits" or "free buffer waits" are in the top 5 wait events, which, to me, means that there is no buffer contention.
What I understand is that after enabling DirectIO, it takes longer for the IO to complete because pre-DirectIO it would return from the file system cache whereas now it has to return from the disk (and I do see at the OS level that IO has become slow), but should that result in dbwr consuming more CPU?
Infact after enabling DirectIO, IO has become very slow which is another problem and as a result log file writes have also become slow which is a 3rd problem. btw, I am aware that if there were many FTS, then DirectIO can make the system slow but there are no FTS in my case. Also, that SGA should be increased after enabling DirectIO, which has also been done.
Thanks

user12022918 wrote:
DBWR is consuming high CPU. After enabling DirectIO on Solaris SPARC 10, dbwr is eating away almost 1 CPU on a v440 machine i.e. 19% throughout the day. 19% is less than 1/5th of a CPU. Or are you referring to a 100% being all 4 CPUs?
What I understand is that after enabling DirectIO, it takes longer for the IO to complete because pre-DirectIO it would return from the file system cache whereas now it has to return from the disk Incorrect. See directio for details.
Yes, removing the file system cache from the I/O layer for a device can reduce I/O performance if the caller does not perform its own caching. However, direct I/O will eliminate the system cache overheads (and associated CPU resources needed) from a caller (like Oracle) that implements its own sophisticated buffer cache.
Direct I/O should therefore increase Oracle I/O performance and decrease resource footprint as it eliminates the need for the kernel to maintain a cache for that device.
Infact after enabling DirectIO, IO has become very slow which is another problem and as a result log file writes have also become slow which is a 3rd problem. Direct I/O, as per the Sun docs, is an advisory call. It may not place that device in direct I/O modes. It may result in partial direct I/O. So you need to make sure exactly what happens and how successful (partial or complete) this setting was.
btw, I am aware that if there were many FTS, then DirectIO can make the system slow but there are no FTS in my case. FTS (multi block reads/large sequential reads) is slower? This is contrary to Sun's docs that state:
Large sequential I/O generally performs best with DIRECTIO_ON, except when a file is sparse or is being extended and is opened with O_SYNC or O_DSYNC.

Gamin/gam_server eating 70%+ of my CPU

Similar Messages

Maybe you are looking for