Welcome to the MacNN Forums.

closer · Feb 20, 2007, 11:47 AM

Hi... I've posted in other forums on this problem but have not been able to get a decent reply... just found this forum and it looks pretty good so here goes...

I'm using a G4 533Mhz 1G ram + additional NIC as an internet web proxy server at school

dansguardian/Squid as the proxy SW

it rarely stay up and running for more than 8 minutes now... in the past week I've

• UPdated sys SW to 10.3.9
• switched out the RAM completely and replaced with new RAM
• Run Disk Utility and repaired permissions
• Booted from CD and Run Disk Utility and Run Repair Drive
• Run Tech Tool Pro
• Run Drive 10
• Run Disk Warrior
• swapped out the HD with a new one
• run without the NIC
• tried different RAM slots

that kinda lead me to believe that it might not be the HD failing... it appears to check out ok on all four Utility programs... and the sys SW has been fully updated... and the fan appears to be working fine so things shouldn't be overheating and causes freezes and crashes...

I am running no other programs on this server... it just won't stay on very long, and my idea of a job is not restarting a computer and proxy server every 8 minutes... dansguardian and Squid are set to reboot automatically on restart... but they only do that half the time... otherwise its a manual terminal boot for them

any ideas why this mac has become so disfunctional... I'm not particularily Unix savvy so please elaborate if you're talking Unix speak

the only thing I can think of now is that it might be the CPU... in which case I have another disfunctional G4 here... is it possible to swap the CPU from that one into this one...

or could it be a lazy power supply... I could probably swap the one from the other mac in that case too...

any help you can provide would be greatly appreciated... the school is starting to go nuts on me... no internet!

thanks, dan

reader50 · Feb 20, 2007, 12:48 PM

Originally Posted by closer

...
dansguardian/Squid as the proxy SW

it rarely stay up and running for more than 8 minutes now...

What exactly happens when "it" quits. Do the two server software programs quit? Does the PowerMac shut down, or the system software kernel panics?

If it's the server software programs, you need to check their logfiles, see what errors got logged right before they shut down.

If 10.3 locks up, check the console log in /Applications/Utilities/Console.app

If 10.3 kernel panics, use the Console utility above. Click the Logs tab, the use it to check for /Library/Logs/panic.log. That will be a copy of the panic text, which will likely give a good hint of where the problem is.

If the PowerMac spontaneously shuts down, it's most likely a hardware problem. But the console log or panic log are still good places to check first.

closer · Feb 20, 2007, 01:09 PM

Originally Posted by reader50

What exactly happens when "it" quits. Do the two server software programs quit? Does the PowerMac shut down, or the system software kernel panics?

pretty much yes to everything...

Originally Posted by reader50

If it's the server software programs, you need to check their logfiles, see what errors got logged right before they shut down.

If 10.3 locks up, check the console log in /Applications/Utilities/Console.app

If 10.3 kernel panics, use the Console utility above. Click the Logs tab, the use it to check for /Library/Logs/panic.log. That will be a copy of the panic text, which will likely give a good hint of where the problem is.

If the PowerMac spontaneously shuts down, it's most likely a hardware problem. But the console log or panic log are still good places to check first.

I'm kinda leaning to this one... it can crash at various stages of startup, crash during login, crash while in terminal, crash while in console... and while dansguardian and squid are startup items... they don't always restart upon restart... thats why I'm kinda thinking more HW than SW problem... cuz even with the newly formatted drive in place of the old HD it froze... so dansguardian and Squid weren't the issue there!

this is probably a stupid question but is there any way this could be lithium battery/PMU problem?

I've checked log files til they're coming out the yingyang and most come down to Memory access exception... and most freezes and restarts are right after
/usr/share/servermgrd/cgi-bin/servermgr_web: [517 - a000b2a4] SIGALRM: process timed out
and repeats this until it freezes and restarts

does that ring a bell anywhere?

closer · Feb 20, 2007, 01:27 PM

kernel panic log

unresolved kernel trap(cpu 0)
-- then a bunch of numbers etc
backtrace terminated - unaligned frame address: 0xFE05C821

reader50 · Feb 20, 2007, 01:48 PM

Give us the text of the KP log. Or the last 30 lines or so, if it's super long.

So it:
• sometimes keeps running, but the server programs die
• sometimes 10.3 freezes
• sometimes 10.3 kernel panics
• sometimes the PowerMac spontaneously shuts down

Is that about right, on all points? A bad OSX install could cause the first three, but the last one screams hardware fault.

Note that a bad HD (or HD controller) can cause startup problems and freezes, but won't cause the system to power down. The same applies to a bad graphics card. Or even bad RAM.

If it does power down sometimes, check the power cord and socket. Bad cord / loose connection / loose power socket / burned power socket. Any of those would cause all the problems you list, and be fairly cheap to fix. If in doubt, plug an incandescent desk lamp into the same power source, see if it flickers. That would indicate a problem in the wall socket, or further upstream.

P · Feb 21, 2007, 04:25 AM

If it goes down rather quickly after starting up, it could be leaking capacitors on either the motherboard or in the PSU. Open it up and check visually if you see anything bulging.

closer · Feb 21, 2007, 08:05 AM

Originally Posted by reader50

Give us the text of the KP log. Or the last 30 lines or so, if it's super long.
So it:
• sometimes keeps running, but the server programs die
• sometimes 10.3 freezes
• sometimes 10.3 kernel panics
• sometimes the PowerMac spontaneously shuts down
Is that about right, on all points? A bad OSX install could cause the first three, but the last one screams hardware fault.

if I can keep the mac up long enough to copy the file to a USB disk I'll send it through... or not... can't see where to attach a file in this forum... regardless, this morning I changed the lithium and reset the PMU and have already frozen 4 times during early boot process and once at login screen

as per 4 points... seldom stays up and running now... and will freeze under any circumstance... it only restarts because it has been set up to do a restart on freeze... it actually doesn't just shut down and stay down...... so last point does not really apply

Originally Posted by reader50

Note that a bad HD (or HD controller) can cause startup problems and freezes, but won't cause the system to power down. The same applies to a bad graphics card. Or even bad RAM.

I've switched out RAM completely and replaced with new RAM... I've started from an external firewire drive with 10.3.9 installed and it still freezes... I have another G4 with an identical graphics card which I will test this morning... although there have been no instances of video problems to date...

Originally Posted by reader50

If it does power down sometimes, check the power cord and socket. Bad cord / loose connection / loose power socket / burned power socket. Any of those would cause all the problems you list, and be fairly cheap to fix. If in doubt, plug an incandescent desk lamp into the same power source, see if it flickers. That would indicate a problem in the wall socket, or further upstream.

it is hooked up to a Smart-UPS 1000XL with battery pack to absorb our weekly power hits...

doesn't look like I'm going to get this baby up and running to be able to send the most current KP file... I do have a system.log available if any has a burning desire to plod through that ... its usually thread 1 that has the errors

ok... late breaking news...server up, console running, error logs saved... panic.log, system.log, ApplePasswordServer.log... are any of these useful to you?... and let me know how to attach a file and I'll post them next post

closer · 08:28 AM

duplicate post cancelled... inserting KP text here instead... forget what I said about not being able to edit... it just reappeared! go figure!... also the date appears as March in the log below... it is todays date... date was reset after the restart due to PMU reset... ie... I haven't time travelled!
most recent KP panic.log
*********
Tue Feb 20 14:40:34 2007
Unresolved kernel trap(cpu 0): 0x700 - Program DAR=0x0000000029C2B000 PC=0x0000000000003188
Latest crash info for cpu 0:
Exception state (sv=0x2C8D6000)
PC=0x00003188; MSR=0x02083000; DAR=0x29C2B000; DSISR=0x42000000; LR=0x0000256C; R1=0xFE05C821; XCP=0x0000001C (0x700 - Program)
Backtrace:
backtrace terminated - unaligned frame address: 0xFE05C821
Proceeding back via exception chain:
Exception state (sv=0x2C8D6000)
previously dumped as "Latest" state. skipping...
Kernel version:
Darwin Kernel Version 7.9.0:
Wed Mar 30 20:11:17 PST 2005; root:xnu/xnu-517.12.7.obj~1/RELEASE_PPC
panic(cpu 0): 0x700 - Program
Latest stack backtrace for cpu 0:
Backtrace:
0x00083498 0x0008397C 0x0001EDA4 0x00090C38 0x0009402C
Proceeding back via exception chain:
Exception state (sv=0x2C8D6000)
PC=0x00003188; MSR=0x02083000; DAR=0x29C2B000; DSISR=0x42000000; LR=0x0000256C; R1=0xFE05C821; XCP=0x0000001C (0x700 - Program)
Backtrace:
backtrace terminated - unaligned frame address: 0xFE05C821
Kernel version:
Darwin Kernel Version 7.9.0:
Wed Mar 30 20:11:17 PST 2005; root:xnu/xnu-517.12.7.obj~1/RELEASE_PPC
*********

closer · Feb 21, 2007, 08:22 AM

sorry about the double post... our internet service is very much pooched right now... trying to go through hispeed off pooched server when its up for a fem minutes and very slow board WANspeed... Bd server told me it didn't post... anyway tried to edit it out but now don't even have those options... this is wonderful!

P... no capacitors are bulging... but thanks for the tip anyway

hope this doesn't double-post too... hispeed only lasted a few minutes this time... seems like as soon as I try to make contact with the alien thing (pooched G4) it freezes... does this not spell CPU trouble... and if so... can I simply swap this one out with the CPU from my other G4... they are both 533's

Leonard · Feb 21, 2007, 09:04 AM

Do you have any USB devices attached to it other than the mouse and keyboard? You mentioned a UPS, does it have a USB connection to the PM (PowerMac)? Some do.

Actually, the environment doesn't sound too computer friendly either, you talk about "weekly power hits". Maybe you need to take the PM out of that environment and try it or maybe the environment has already damaged the PM.

closer · Feb 21, 2007, 12:04 PM

Originally Posted by Leonard

Do you have any USB devices attached to it other than the mouse and keyboard? You mentioned a UPS, does it have a USB connection to the PM (PowerMac)? Some do.

Actually, the environment doesn't sound too computer friendly either, you talk about "weekly power hits". Maybe you need to take the PM out of that environment and try it or maybe the environment has already damaged the PM.

nope no other USB devices hooked up... and I need this G4 here to run our proxy server... secure facility that requires firewall and monitoring, therefore two cards... at least until we can update our G5 server to 10.4.. I understand I can find a workaround through 10.4 somehow...

I also ran hardware test again and everything checks out ok

also tried Rember (couldn't get memtest and was told Rember is the GUI equivalent)... but it froze while doing this test too... prior to freeze itlooked like it was doing... nothing... no activity on status bar at all... Rember works great on my dual intel iMac workstation... but nothing happens on the G4... does this mean I'm getting somewhere? ha, ha

reader50 · 06:47 PM

Originally Posted by closer

...
I've switched out RAM completely and replaced with new RAM... I've started from an external firewire drive with 10.3.9 installed and it still freezes... I have another G4 with an identical graphics card which I will test this morning... although there have been no instances of video problems to date...

You've eliminated the RAM as a possible issue. Also the HD controller, and the OS install, assuming the OS copy on the FW drive was a decent install.

The only remaining thought on the OS install, is what disk was used to create the installs? A G4 533 did not come with an OSX Install CD, so it's either a Retail Panther CD, or the Install/Reinstall CD from a later Mac model. You want the Retail disk, the install CDs from other Mac models often omit drivers for other models.

Originally Posted by closer

it is hooked up to a Smart-UPS 1000XL with battery pack to absorb our weekly power hits...

Try plugging in directly to the wall, bypass the UPS for awhile. If the UPS has gone bad, that would explain things too. And check the PM's power cord anyway. I once diagnosed a new sunflower iMac, which had issues if the case were nudged in any way. Like plugging a USB cord into it. Turns out the power cord wasn't fully plugged into it, nudges interrupted the power input. After pushing the cord all the way in, it worked fine. If your cord feels loose in any way, replace it.

Originally Posted by closer

ok... late breaking news...server up, console running, error logs saved... panic.log, system.log, ApplePasswordServer.log... are any of these useful to you?... and let me know how to attach a file and I'll post them next post

The system.log would be interesting, except that your startup hangs at different places. Host a copy of the file somewhere, and I'll look at it. But this is a long shot, since the hangs happen at different startup points.

Originally Posted by closer

*********
Tue Feb 20 14:40:34 2007
Unresolved kernel trap(cpu 0): 0x700 - Program DAR=0x0000000029C2B000 PC=0x0000000000003188
Latest crash info for cpu 0:
- snip -
Backtrace:
backtrace terminated - unaligned frame address: 0xFE05C821
Proceeding back via exception chain:
Exception state (sv=0x2C8D6000)
previously dumped as "Latest" state. skipping...
- snip -
Backtrace:
backtrace terminated - unaligned frame address: 0xFE05C821

Normally, the backtrace and exception chain lead back to the driver(s) that caused the crash. This would be the case if it's an external device (external to the motherboard) such as a USB, FW, or ATA device. Or if it's a software issue. Having the backtraces fail, suggests program execution was interrupted in mid-code by a hardware issue. Go ahead and try swapping the graphics card, but it's looking like the graphics card is not a suspect after all.

It looks to me like:
• Power source problem (PowerMac's power cord, or UPS going bad)
• Weak / old power supply unit in the PowerMac
• Bad / overheating CPU
• Bad motherboard

The first one would be easy to fix, swap cords or UPS. Test that item by plugging directly into the wall for awhile, preferably with a different cord than usual on the PowerMac.

The 2nd one can be tested for by swapping in a different PSU, or by reading output voltages from the existing PSU. See if the outputs (which should be +/-12V and +/-5V) are dropping well below what they should be. Testing this would require more skills and/or tools, but a reasonably handy person could do it.

The 3rd can be tested if you have a spare 533 CPU from an identical G4. You'd need to scrape off the old thermal paste from the CPUs & heat sinks, then apply a small amount of new thermal paste during the reassembly process. If you get this step wrong, the replacement CPU will burn up.

The 4th you should stay away from, moving motherboards between PowerMacs is more involved.

Since this is a work G4, which you need to have up and running, I'd suggest swapping in a spare PowerMac. Move the HD, the new RAM, and any other essential components (the 2nd NIC?) over to the spare PowerMac. Get your system up and running, then further diagnose the problem G4 if you like.

Note that a G4 533 today is only worth a couple hundred at most, you probably have more labor than that in it trying to fix it. Putting more major labor or money into it isn't worth it.

closer · Feb 21, 2007, 12:35 PM

2 progs that I've found TheXLab.com and looked like they might divulge something were Rember and MacSysInfo

tried both out on my own workstation and they work great... but neither function on the G4... both are versions that supposedly work with 10.3.9

console error messages were

Rember[444] *** -[NSCFArray addObject:] attempt to insert nil

and

MacSysInfo[448] Unable to load nib file: MainMenu, exiting

closer · Feb 21, 2007, 12:51 PM

Hi reader50

first of all... I really appreciate your assistance here... I'm actually starting to think that this aggravation is now amounting to more than its worth... have switched the power cord... the 2 little progs still don't work, but haven't switched to a different power source yet... one thing at a time!

my sys.log is hosted at another forum (sorry) in post#14...
http://forums.macosxhints.com/showthread.php?t=68026

don't know if that will be of any help or not... and yes I do have another power supply available (in a G4 that also hasn't been functional for awhile!!!) but at least there is one there... no harm in trying I guess... not sure i know anything about the thermal paste though... I'd probably end up toasting the CPU as a result... but I have a CPU available well if needed... I'll try anything...

olePigeon · Feb 21, 2007, 01:49 PM

Try resocketing the CPU daughter card. Also, try booting off only 1 stick of RAM. Take all the other RAM out. Try replacing the ATA/IDE cable with a different one, and try putting your main HDD on the 2nd ATA BUS.

closer · Feb 21, 2007, 01:50 PM

ok... so unless I had two bad power supplies... it wasn't the power supply or fan either... replaced both... and froze on startup... have rebooted successfully but who knows for how long! also didn't look like it did a full boot through all the startup processess... maybe a few then right to login screen

closer · Feb 21, 2007, 01:53 PM

Originally Posted by olePigeon

Try resocketing the CPU daughter card. Also, try booting off only 1 stick of RAM. Take all the other RAM out. Try replacing the ATA/IDE cable with a different one, and try putting your main HDD on the 2nd ATA BUS.

oh man... this will sound bad... but where is the "daughter" card?

have already tried booting from single RAM and from various slots... haven't tried the alternate HD cable yet though

reader50 · 06:50 PM

As expected, the system log doesn't show anything interesting right before each freeze. Afterwards, it blames the restart on the PMU, but that likely means you hit the reset button to force a reboot.

The CPU daughter card is the card the CPU is soldered to. It's under the heat sink. The CPU doesn't directly plug into a motherboard socket, instead the CPU daughtercard plugs into the motherboard. It would be worthwhile to try reseating it, except I'm not sure if you can unplug it without removing the heat sink first. If you do that, then you need to clean off & replace the thermal paste, same as for a CPU swap.

If it's anything like my Sawtooth's original heat sink setup, then you need to release the clips holding the heat sink to the CPU. Pull off the heat sink, which breaks the old thermal paste. The screws that lock the daughtercard to the motherboard will be visible now, you couldn't see them (much less unscrew them) until the heat sink is removed. Don't try this unless you have replacement thermal paste handy.

It might be worth it to do the RAM test, using one stick in different slots - if you haven't tried that already. RAM slots have been known to go bad a few times, but it's pretty rare. You've essentially eliminated the RAM itself, by replacing it with new.

I doubt the ATA cable will make any difference, since you booted from a FW drive, and had the same symptoms. Ditto for putting the internal HD on the 2nd ATA bus, which is the one supplying the optical drive. That would amount to moving the regular HD to the Zip bay. It's actually easy to do that, so try it if you like.

The log has a few complaints about block allocations on the HD, but that is almost certainly caused by all the freezes. If you can get it to stop freezing, Disk Warrior will clean all those up. Those are symptoms, not the cause.

Dansguard doesn't start up if Squid fails, since it requires Squid. Squid sometimes fails to start up, because a child process didn't start up properly. This might be a timeout issue, caused by the frequent restarts - ie, if restarts happen more quickly than x minutes, squid can't restart clean - this might be the case, if it refers to something on the internet (an authorization service?) and the internet side requires a minimum time between references. I don't know anything about squid, so this is pure speculation.

"/usr/share/servermgrd/cgi-bin/servermgr_web: [517 - a000b2a4] SIGALRM: process timed out"
I don't know what servermgrd is (server manager daemon), looks like it has trouble starting up. However, the symptoms suggest a hardware problem rather than software, so I'd be inclined to chalk this error up to frequent restarts & minor HD damage too. In other words, another symptom rather than a cause.

You've replaced the PSU with no change. Did you bypass the UPS yet? If the UPS is eliminated and the power cord too, that leaves the CPU and the motherboard. Unless one of the long shots works (IDE cable, moving internal HD to the Zip bay).

Edit: it's interesting that you were able to boot from the hardware test CD and it didn't freeze. Try booting the OSX Install CD, then leave it alone for awhile. Or jump in and out of Disk Utility / Terminal a few times. The Install CD does boot up the majority of OSX. If the system is perfectly stable on the Install CD, then your OSX Install would start looking suspicious again. Ditto for the OSX copy installed on the FW drive...

closer · Feb 22, 2007, 07:34 AM

once again R50, thank you for your extensive answer... sounds like I've been able to cover most of the possibilties...

came in early this morning and swapped the HD into the other G4, although the power button wouldn't stay lit (I then pushed the PMU button once and it came on nicely the second time round)... it is on now and I'll be testing it out this morning... if this doesn't work then you'll probably hear me screaming from all the way out east...

anyway, I do appreciate your help and will let you know how we make out... if it works it was clearly a MB or CPU problem

closer · Feb 22, 2007, 11:27 AM

afternoon update...

after making the HD switch to the alternate G4 the world seems to be in order again... (touching real wood here) the server has stayed up and running since the first boot this morning... so it really looks like it was either CPU or MB related after all...

guess the best bet then is to salvage parts from the first G4 for later use if needed on the usable one now?... R50, as you suggested earlier there probably isn't much sense spending any more time or money on a mac this old... cost of repair or replacement of MB/CPU wouldn't make sense...

anyway... I really appreciate your assistance with troubleshooting the problem I had with this mac... and I do believe nnforums has just become my mac forum of choice