Mitchell Posted July 27, 2007 Report Share Posted July 27, 2007 Hi people. Have been using Mandriva 2007.1 for a month now, with Windows XP installed on Virtual Box for my business database, haven't worked out how to move that over yet. Anyway, just started having problems with freezing. Had a stable install, then managed to stuff it up. I was installing lots of packages from the software list, and somewhere along the way I think I installed something that stuffed everything up. My box was freezing in the middle of using it, and it sucked. Luckily I've been making backups with Mondo, so restored everything to an earlier time. Everything was fine at first, then my system started freezing after being left idle for awhile, generally a couple hours. Each time it freezes I've had to manually reboot it, no keyboard shortcuts work. I have restored to several earlier times now, including the first Mondo backup I made, right after a clean install. I wiped both my Root "/" and usr "usr" partitions, reformated them with Mondo, and restored the system data from my first mondo backup. Left my box on all night last night to check it, and this morning it was frozen on the default Mandriva screensaver. (this has never worked properly incidentally, always has a thick orange border on both sides. Usually don't use it) I have not touched my home "home" partition, and I also have a "guestos" partition where windows is installed under Virtual Box, and a "backup" partition where Mondo has been saving my backup Iso images. None of these partitions have been changed, and I'm hoping not to have to. Having just switched to Mandriva, my wife isn't very happy, and will really hate it if I have to reinstall and she loses all of her emails. She's not very computer savvy, and I'm trying to make the shift as painless as possible for her. Anyway, the odd thing about all this is that it has been working fine up to about a week ago. Any ideas on what's wrong? Everything else is working fine, although my sound driver seems to periodically disable itself, and I need to reboot the system. I haven't tried a reinstall of the "usr" and "/" partitions, and I hope I won't have to. Still trying to get past that Microsoft mentality. Is there any chance there is some setting in my "home" partition that is causing this? What could I look for? Thanks for any help, ~Mitchell Quote Link to comment Share on other sites More sharing options...
arctic Posted July 27, 2007 Report Share Posted July 27, 2007 Hi there. :) First of all: Your /home partition is very, very unlikely to cause the freezing (chances are 1:1000). As long as you have everything on a separate /home partition, no emails and whatsoever will be lost. It is all safe, even if you reinstall one day. That being said, I know how you feel. I had the same freezing problem on my desktop system and in my case it was caused by the gdm login manager not interacting very well with the x-server, which was very difficult to find out in my case. Once I used the init 3 mode for logging in, there were no more freezes in my case. In your case, I would start to narrow down the source of your problems by -checking the /home/username/.xsession-errors file -checking the /var/log/messages file. There, check the last 50 or so lines from a date when the system froze. Maybe there is a hint. If you cannot decipher it, upload your messages file to e.g. pastebin ( http://pastebin.com/ ) -post the hardware you use. Maybe there is a known troublemaker included in the list. Anyway: Good luck. Quote Link to comment Share on other sites More sharing options...
Mitchell Posted July 27, 2007 Author Report Share Posted July 27, 2007 Thank you so much, I'll try all those things. I really hope you're right about the /home partition, it will be a real hassle if I ever have to reformat it. It's odd though, I can't think why when I restore the / and /usr partitions to a time when everything was working why I'm still having this problem. Thanks Arctic, will let you know how it goes. ~Mitchell Quote Link to comment Share on other sites More sharing options...
Ixthusdan Posted July 27, 2007 Report Share Posted July 27, 2007 Do not rule out hardware. If your RAM and processor test well, then the most likely candidate is your power supply. Quote Link to comment Share on other sites More sharing options...
AussieJohn Posted July 28, 2007 Report Share Posted July 28, 2007 As Lxthusdan says. Another very strong likelihood is memory. Actually I found this to be more often the case than anything else. If you have multiple sticks of matched ram (You shouldn't really be using mixed sticks of ram anyway) then over time one stick may no longer match the other and may even be on its way out. I have solved one incidence (current condition) by going into the bios and setting a next lower memory speed and have had no re-occurrence for over a year now, even though the earlier settings were exactly as recommended by the manufacturer. Cheers. John. Quote Link to comment Share on other sites More sharing options...
Mitchell Posted July 29, 2007 Author Report Share Posted July 29, 2007 (edited) Hardware, what a bugger...again. Had this system built from scratch a couple years ago, and was lucky to buy the 3 year extended warranty. I'm so glad I did. A short list of what's needed to be replaced? 80 gig seagate harddrive, cd-rom drive (twice), and power supply (twice). Currently one of the usb drives has cacked it, and I've been wondering what's next. Oddly enough everything to date (save the USB) has been covered by the warranty, so apart from the hassle I haven't been that fussed. The nearest thing to an explanation from a repair technician (unassociated with the company I bought the box from) is that parts are made so fast and cheaply these days that this is the result. Can anyone confirm this? Have a feeling I may have bought cheaper parts, and had someone who didn't know that much about computer's put them together... and the store manager seemed so knowledgeable at the time! Anyway, JOHN: you said it's not a good idea to mix memory sticks. Why? I've got a single 512 mb stick of ddr2 ram, and have been thinking about buying either another 512, or 1G to go with it. Is this not a good idea? My computer in general runs quite fast, unless I'm doing something like using Virtual Box to run Windows. (which I do quite alot actually, still haven't extracted my business database from MS Windows yet.) Also, TXTHUSDEN, how do I check my RAM and processor? Is there a way I can test this and my power supply without buying anything new. The extended warranty should still cover me, but running Linux I've a feeling I'll be told all this is a software problem. Anyway to confirm all this? Thanks again for all your help. ~Mitchell Edited July 29, 2007 by Mitchell Quote Link to comment Share on other sites More sharing options...
AussieJohn Posted July 29, 2007 Report Share Posted July 29, 2007 Hello Mitchell. All the hardware manufacturers, especially memory makers recommend against mixing memory sticks. You can add another 512 or even two more of 512 but they should be matching types to your existing one. If you add a 1Gb to the mix then you can get into trouble unless all the sticks are 1Gb and of the same type. Sure you can mix them and maybe they will work but it will only be a while before you start to have troubles. I can speak from first hand experience because in earlier years I used to do such things until I learned to do things differently. I have also learned it is better to have dual sticks of memory instead of only one larger one. If one fails, even if it is covered by a lifetime warranty (good quality stuff) you still have half of your memory to continue to work with. If you are looking at 1Gb stick why not consider two matching 1Gb sticks such as Corsair do with their TwinX series of matched pairs. Cost of memory is relatively cheap nowdays and Corsair are excellent value and quality. What wattage is your Power Supply ?? It is possible that your builder has put in a 250watt unit. (a cheap minimum) You should be using 350w or higher so that the power supply is cruising instead of being run close to its limits. I use a 450w and have not ever had a power supply problem with it in the past 3yrs. Cheers. John. Quote Link to comment Share on other sites More sharing options...
arctic Posted July 29, 2007 Report Share Posted July 29, 2007 In order to check your RAM, you can use a live-CD (like e.g. knnoppix) and run memtest86 from the cd-bootmenu. Quote Link to comment Share on other sites More sharing options...
daniewicz Posted July 29, 2007 Report Share Posted July 29, 2007 Have a feeling I may have bought cheaper parts, and had someone who didn't know that much about computer's put them together... and the store manager seemed so knowledgeable at the time! If you have replaced your power supply and CDROM twice, then you are dealing with low quality hardware. You definitely need to check your RAM as arctic says. Quote Link to comment Share on other sites More sharing options...
Greg2 Posted July 29, 2007 Report Share Posted July 29, 2007 I'll have to add to this, that when your system freezes the next time; open the case and see if all the fans are running, power supply fan, cpu fan, gpu fan, case fan, etc.? Also check to see if the hard drive light is on, like it's trying to be accessed? I've recently replaced a mother board with a faulty controller that showed this symptom. Quote Link to comment Share on other sites More sharing options...
Mitchell Posted July 31, 2007 Author Report Share Posted July 31, 2007 Thanks everyone, still working through all this. Still freezing, though only when I haven't used the computer for a few hours. Also, when I reboot my system, my sound doesn't work. I click on the sound icon in my top panel, and all I get is a microhone. When I reboot a second time everything is fine, working, and seems really fast. Thanks for the advice about low quality hardware, I think I'm learning my lesson. I'm still hoping this is a software issue..but... ARCTIC, can I also use the Mandriva 2007.1 live cd to check RAM? (this is what I originally installed my system from) Also just looked in my messages file, under /var/log. Noticed today that it seemed to freeze right after restarting my ethernet wireless controller... Am going to disable this and leave the box on overnight. If it doesn't freeze problem solved...at least until I need to use my wireless card again...lol. I'm using a Belkin wireless Desktop card in addition to my usual ethernet connection cabled to the router. Basically I was networking my Desktop and Laptop pc's through my Belkin router in microsoft windows, and haven't sorted that out under linux yet. So I've been playing around with my connections, that might be part of the problem. Anyway, will let you know. Thanks for all the tips so far. Quote Link to comment Share on other sites More sharing options...
daniewicz Posted July 31, 2007 Report Share Posted July 31, 2007 You can check your RAM and do other diagnostics with the Ultimate Boot CD http://www.ultimatebootcd.com/download.html Quote Link to comment Share on other sites More sharing options...
Greg2 Posted July 31, 2007 Report Share Posted July 31, 2007 can I also use the Mandriva 2007.1 live cd to check RAM? You could just do urpmi memtest86+ reboot, and select it from the boot menu. Quote Link to comment Share on other sites More sharing options...
Mitchell Posted August 1, 2007 Author Report Share Posted August 1, 2007 (edited) Installed memtest86+ , had no trouble running it from the boot menu. Unsure how it works, i.e. I let it tun for 2 hours and it was still running tests... got to test 9 and I rebooted, it just seemed to be looping itself and running the same test over again. Did I miss something? Will it eventually finish, and is two hours a standard time for it to run without finishing? Am posting the log of my message file ( /var/log/messages ) where my computer has frozen the past two times, both last night and yesterday. Last night I turned off the radio for my wireless card, but I couldn't disable it. Have been playing around with Mandriva Control Centre today, and all I've managed has been to get the icon for internet connection to turn into green bars, the wireless icon. Short of pulling out my wireless card, or finding the driver for this and removing it, are there any other ways I can temporarily disable this to see if this is my problem? Have been able to untick it as an active interface under the connection icon, but am unable to actually disable it. Am unsure how to uninstall this also... Thanks. (copy of /var/log/messages, up to when I restarted my box after the last freeze) Aug 1 02:12:09 localhost kernel: SoftMAC: Open Authentication completed with 00:11:50:8c:be:58 Aug 1 02:19:48 localhost kernel: eth0: Media Link Off Aug 1 02:19:49 localhost ifplugd(eth0)[4183]: Link beat lost. Aug 1 02:19:50 localhost ifplugd(eth0)[4183]: Link beat detected. Aug 1 02:19:53 localhost kernel: eth0: Media Link On 100mbps full-duplex Aug 1 02:19:54 localhost kernel: eth0: Media Link Off Aug 1 02:19:55 localhost ifplugd(eth0)[4183]: Link beat lost. Aug 1 02:19:56 localhost ifplugd(eth0)[4183]: Link beat detected. Aug 1 02:19:59 localhost kernel: eth0: Media Link On 100mbps full-duplex Aug 1 02:33:27 localhost kernel: martian destination 0.0.0.0 from 192.168.1.1, dev eth0 Aug 1 02:54:10 localhost kernel: SoftMAC: Open Authentication completed with 00:11:50:8c:be:58 Aug 1 03:01:01 localhost crond[23467]: (root) CMD (nice -n 19 run-parts --report /etc/cron.hourly) Aug 1 03:36:12 localhost kernel: SoftMAC: Open Authentication completed with 00:11:50:8c:be:58 Aug 1 09:25:56 localhost syslogd 1.4.1: restart. (prior var/log.messgaes file, freezing time before last) Jul 31 14:18:29 localhost kernel: SoftMAC: Open Authentication completed with 00:11:50:8c:be:58 Jul 31 14:26:56 localhost ntpd[6528]: synchronized to 203.82.209.217, stratum 2 Jul 31 14:37:15 localhost kernel: martian destination 0.0.0.0 from 192.168.1.1, dev eth0 Jul 31 14:50:15 localhost kernel: NETDEV WATCHDOG: eth1: transmit timed out Jul 31 14:50:15 localhost kernel: bcm43xx: Controller RESET (TX timeout) ... Jul 31 14:50:15 localhost kernel: ACPI: PCI interrupt for device 0000:00:07.0 disabled Jul 31 14:50:15 localhost kernel: ACPI: PCI Interrupt 0000:00:07.0[A] -> GSI 18 (level, low) -> IRQ 22 Jul 31 14:50:15 localhost kernel: bcm43xx: Chip ID 0x4306, rev 0x3 Jul 31 14:50:15 localhost kernel: bcm43xx: Number of cores: 5 Jul 31 14:50:15 localhost kernel: bcm43xx: Core 0: ID 0x800, rev 0x4, vendor 0x4243, enabled Jul 31 14:50:15 localhost kernel: bcm43xx: Core 1: ID 0x812, rev 0x5, vendor 0x4243, disabled Jul 31 14:50:15 localhost kernel: bcm43xx: Core 2: ID 0x80d, rev 0x2, vendor 0x4243, enabled Jul 31 14:50:15 localhost kernel: bcm43xx: Core 3: ID 0x807, rev 0x2, vendor 0x4243, disabled Jul 31 14:50:15 localhost kernel: bcm43xx: Core 4: ID 0x804, rev 0x9, vendor 0x4243, enabled Jul 31 14:50:15 localhost kernel: bcm43xx: PHY connected Jul 31 14:50:15 localhost kernel: bcm43xx: Detected PHY: Version: 2, Type 2, Revision 2 Jul 31 14:50:15 localhost kernel: bcm43xx: Detected Radio: ID: 2205017f (Manuf: 17f Ver: 2050 Rev: 2) Jul 31 14:50:15 localhost kernel: bcm43xx: Radio turned off Jul 31 14:50:15 localhost kernel: bcm43xx: Radio turned off Jul 31 14:50:15 localhost kernel: bcm43xx: Controller restarted Jul 31 18:29:27 localhost syslogd 1.4.1: restart. Edited August 1, 2007 by Mitchell Quote Link to comment Share on other sites More sharing options...
AussieJohn Posted August 1, 2007 Report Share Posted August 1, 2007 Hello Mitchell. The usual recommended time is at least 12 hours or if possible, 24hrs. The longer the better. The tests may seemed to be simply looped but from what I have observed the earlier tests are relatively simple but progressively get tougher and tougher. Each test seems to be repeated a number of times before it moves on to the next test. When I have run tests and had problems indicated, they were usually towards the end of long runs (in my cases, 24hrs). Cheers. John. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.