RHEL 6.5 performance issue
 

I'm new here and this is my first post, so if it doesn't belong in this area please move to correct place.

I'm definitely a linux newbie, my company issued us laptops with RHEL 6.4 on them which took quite a bit of getting used to. It had been working fine until they pushed an update to 6.5 recently. I backed up all my data and allowed the update to proceed and it completed successfully. Since then, I've noticed a serious degredation in startup times for any applications that run under Linux, to the tune of several minutes for most before they are up and usuable.

I'm looking for a way that I can try to determine what's slowing things down, ie., something other than system monitor that may help pinpoint what's going on. I've done some searches for performance issues with this version of Red Hat and come up empty. Any help or advice would be greatly appreciated, keeping in mind that my linux knowledge is very limited.

Since your company has paid for RHEL wouldn't you be better off taking advantage of RHEL's own support service? I'd guess they would be the most qualified people to investigate your problem.
jdk

Welcome to the forums!

It could be that your swap is too small. In general for Linux it should be twice the amount of RAM you have. It could also be a lot of other things. I run CentOS 6 fully updated (to whatever version it's on now, don't really pay attention anymore) on my server and have no performance issues, which suggests that it is specific to your system. I would take it to the IT dept. and have them take a look at it. They have physical access to it and can troubleshoot it easier than we can.

I don't think RedHat works like that...you have to have access to the company's account for RedHat, and only someone high up in the IT department should have that.

That seems like an extremely strange policy. You are working for the company and doing company work which requires a properly functioning system. I don't use RHEL but I was under the impression that if you use it then you are paying for support and you are working for a company that has paid for it and you are doing company work. It's strange that they don't support customers who pay for that support. I guess you know more about it than I do.
:confused:
jdk

That's an antiquated "rule" that for the most part has no basis in reality. It depends entirely on the usage of the system. You don't even need a swap at all if you know the system will never run out of RAM.

I generally just do 4 GB, unless it's a small VM with a sole purpose, in which case I may do 500 MB or 1 GB. Or if it's a development system that could potentially run out of RAM due to a bug in some program under development I might do 16 GB. This is all completely independent of the amount of RAM in the machine. Most of my work is done on a CentOS 6 system with 128 GB of RAM and a 4 GB swap, and the swap never even gets used, just stays at "0" all the time. No way would I give that system a 256 GB swap, that would waste a quarter of the drive for no reason.

OP - if you could post the output of "free -m", it would let us know if RAM/swap has anything to do with your problem.

Thanks for the replies folks. I don't have access to Red Hat support, I'm guessing as was stated only certain people higher up the food chain than I am have that access.

I tried going through our so called helpdesk with this, first person I talked to was in level 2 desktop support, their first response was 'it should get better after about 30 days or so'. So what was supposed to happen after 30 days to magically make it better, there was no answer to that. The next person I talked to said it must be a hardware issue and said the best thing to do was reimage the machine. That would leave me having to reinstall all non-default software again and going through configuration all over again as well.

This has left me to try to figure this out on my own, hence my post here. I will provide whatever diagnostics I can to try to figure this thing out. Here is the output of the command requested;

total used free shared buffers cached
Mem: 7753 7496 257 0 41 1015
-/+ buffers/cache: 6439 1314
Swap: 7887 547 7340

Well you are using over 80% of the available RAM, but there's still some left (at least in that output), and swap usage is low, so that shouldn't affect anything. You should keep an eye on it though, >80% RAM usage is generally rare, unless you're running a resource hog on purpose (eg: Windows in a VM).

Nope, not at all. ANYONE from the company in question should be able to call Red Hat support. They may ask you for a subscription number, and if you don't have it, they can verify your subscription. I've done it numerous times. If you're in doubt, and you're having a problem, it should NOT be a problem to contact your boss/bosses-boss, and ask for such information, in order to get your job done. Nothing wrong with that. Add to the fact that your company obviously has something like a satellite server (for pushing out such updates), and it points even further to you actually having RHEL support.
Sounds like your 'help desk' is about as efficient as most others. How, exactly, reinstalling an OS onto broken hardware would somehow FIX that hardware problem is beyond me. And "wait 30 days" is about as useless. Which brings us back to "Call Red Hat support".
Before you do...what kind of laptop is it? CPU? 32 or 64 bit? And while we're happy to try to help you, I think you're in a bad spot here, and would be FAR better off going through the help desk, and kicking things further up the chain in your company. The reason being, is that you have a company-issued laptop, with a company-issued image, with the updates pushed down by the company/IT staff. Anything you do *MAY* violate your IT policies, and land you in trouble. And unless you have root/admin rights on your laptop, chances are you won't be able to fix it anyway, and whatever you DO fix may just get 'unfixed' by rebooting, and having the image 'corrected'.

My hope in coming here was that I might find the fix to be something simple that I could do myself. Whether I would be able to actually implement the fix, I'm not sure, but if not I could at least have something to point our helpless desk in the right direction with. I've already wasted nearly two weeks working with them to no avail.

Laptop info is;
Thinkpad T420, 64-bit cpu.

I understand what you are saying about it being a company issued laptop and the pitfalls associated with that. However, at this point I'm willing to take a chance that I might be able to find a solution for this, with help of course.

The policy is there so that bozos don't log in and buy all sorts of junk with company money. He himself is not paying for the support, the company he works for does. Thus...you fill in the blanks.

OK, good.

Looks like Oracle fixed this problem in their EL6 distro. Maybe you could try and get updates from Oracle. Run these commands, reboot, and tell me if the problem is fixed.

NOTE: for the first command your password won't show up when you type it. That's normal, security thing.

This could be erroneous...

http://stackoverflow.com/questions/1...t-top-does-not

using the NOT !!! redhat updates repo on a redhat system will null and void the support contract

do NOT use Oracle rpm's on redhat !!!!!!

80% ram is BAD
i have ScientificLinux 6.5 on a antique desktop that is 12 years old and has 1 gig ram and as i type this only about 40% of that 1 gig is in use

something is running in the background that is NOT supposed to be

take a look at the output of the "top" command

just type "top " ( NO "" ) into the terminal and hit enter

then watch the output for a min or two to see what is the HOG

Sorry, didn't realize this would void support.