Announcement

Collapse
No announcement yet.

Symptom question

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Symptom question

    I have an Abit KG7-RAID that I have been using for a home server that may have bad capacitors (they look a little bulged). The symptom I'm seeing is backups sometimes fail. I first saw the problem with amanda reporting:

    /home lev 0 FAILED [compress got signal 11, /bin/tar got signal 13]

    but other partitions completing with no problems. I tried to reproduce the problem by running what I think is the same command amanda uses:

    /bin/tar -X /etc/amanda/exclude-list/exclude.txt -cvf - /home | gzip -v -c > /share2/dave/Home.tar.gz

    Sometimes this command completes with no problems, sometimes I get a seg fault and, least frequently, I get a kernel panic. The system is stable otherwise and running a backup is probably the only significant load the system sees. CPU temperature appears to be fairly nominal (110F to 115F) and I gave the system a thorough cleaning to try to eliminate overheating as a possible cause of the problem.

    This system is old enough that I wouldn't have a problem replacing the MB if the capacitors are the problem. On the other hand, the system is more than sufficient for our needs so I don't want to shell out for a new MB unless this really is the problem. Will bad capacitors cause a load related instability?

    Thanks,
    Dave
    Quitely implementing RFC 1925 wherever I go.

    #2
    Re: Symptom question

    well not the guru here
    but Maybe its more likely your psu although if you are using it as a home server
    (is it on a lot of the time? ok since you got a link to it its probably 24/7?)
    possibly both given age

    normally bad caps will cause hangs and reboots but I suppose it does depend on the actual MB.

    not sure of caps used and some are dogs...but if they "look a little bulged" recap is probably a very good idea. (given age and use too)

    what is brand Model is the PSU used in it?

    Welcome, cheers
    You step into the Road, and if you don't keep your feet, there is no knowing where you may be swept off to." Bilbo Baggins ...

    Comment


      #3
      Re: Symptom question

      Originally posted by starfury1
      well not the guru here
      but Maybe its more likely your psu although if you are using it as a home server
      (is it on a lot of the time? ok since you got a link to it its probably 24/7?)
      possibly both given age
      System is on 24x7x365. It's my own little corner of the internet. Other than a few reboots for software updates or powerdowns to replace hardware it's been running continuously since December 2002.
      Originally posted by starfury1
      normally bad caps will cause hangs and reboots but I suppose it does depend on the actual MB.

      not sure of caps used and some are dogs...but if they "look a little bulged" recap is probably a very good idea. (given age and use too)
      I'll probably just replace the MB and PSU since this is also does my web and mail serving. If I need to spend the money, I'll spend it. I just don't want to trash something that's still good because it might be the problem.
      Originally posted by starfury1
      What is brand Model is the PSU used in it?
      The PSU is an Antec SmartPower 400w (yes, I've seen the Antec thread here). MB is an Abit KG7-RAID.
      Originally posted by starfury1
      Welcome, cheers
      Thanks,
      Dave
      Quitely implementing RFC 1925 wherever I go.

      Comment


        #4
        Re: Symptom question

        the antec is the prime culprit here.
        open it up and look at all the bulging/leaking fuhkkyu caps.
        i would do a preemptive recap of the mobo and psu if it were mine.

        Comment


          #5
          Re: Symptom question

          Originally posted by kc8adu
          the antec is the prime culprit here.
          open it up and look at all the bulging/leaking fuhkkyu caps.
          i would do a preemptive recap of the mobo and psu if it were mine.
          I've got a spare PSU someplace. I'll swap it in and see if things improve.

          Sad thing is that I specifically bought Antec PSUs because they had such a good reputation at that time. I think my spare is fairly generic but low run time (got used by my nephew for a couple of weeks when he needed a spare).

          Thanks,
          Dave
          Quitely implementing RFC 1925 wherever I go.

          Comment


            #6
            Re: Symptom question

            By the way the cap company is actually fuhjyyu and it tends to be the biggest problem. A simple quick visual inspection could give you all of the answers that you desire. No need to be afraid of it, just look in at the motherboard see if any of the caps are bulging or are of known bad companies such as GSC or OST etc., open the PSU (usually 4-6 screws and it opens fine) see for bulging caps. That will save you time and let you know what it could be. Rather than switching parts and waiting for an error to occur

            Comment


              #7
              Re: Symptom question

              Originally posted by PedroDaGr8
              By the way the cap company is actually fuhjyyu and it tends to be the biggest problem. A simple quick visual inspection could give you all of the answers that you desire. No need to be afraid of it, just look in at the motherboard see if any of the caps are bulging or are of known bad companies such as GSC or OST etc., open the PSU (usually 4-6 screws and it opens fine) see for bulging caps. That will save you time and let you know what it could be. Rather than switching parts and waiting for an error to occur
              should have put the real name with the common name after it.
              btw the o.p is encouraged to post pics of the mess in that antec and the board.

              Comment


                #8
                Re: Symptom question

                Other than a few reboots for software updates or power downs to replace hardware it's been running continuously since December 2002.
                Thats like 5 years.... 43,800 hours of up time.....not a bad "innings" really considering. (cricket speak)

                so going by the above, its on the way down and will only get worse as time goes on, and the caps age (dry out) more.
                (any electrolytic cap does have a finite life)

                Yeah from what I read around here
                (it does depend I suppose on exact model)

                The recap of the Antec with quality caps does make for an OK supply, is my understanding (its actually "not" a junk supply just bad choice of caps used)
                (stand to be corrected on the above thought)

                I would most defiantly recap the MB and PSU plus rework solder joints in psu heavy sections.
                considering its use.

                but if you are happy to replace both and reload/install all software you could get a better server out of that these days.

                If you are happy with it as it is performance wise in a reliable working condition then recapping is the best alternative.

                Its not that hard but does required a bit of practice to get your soldering skills up to par and of course a soldering iron or 2, read the FAQ's on recapping

                If you don't want to attempt a recap yourself then get a quote on a recap from TC (forum owner) or other approved forum members in your part of the world.

                A local "decent" electronics repair shop should be able to do this work for you but I myself would much prefer to deal with someone that knows and fully appreciates the true nature of MB, PSU recapping

                As KC8 said, crack open you PSU and post some photos of it and the MB too if you don't mind

                As the old saying goes a picture is worth a thousand words.

                BTW may understanding is, it is preferred if you post photos on the forum site rather then off site links.
                (they tend to die and change and your left with txt and no photo)
                but I think thats up to you
                (the site does snuffle them for prosperity thought ).

                Anyway the choice is yours
                please keep us posted on the situation, thanks

                HTH Cheers
                You step into the Road, and if you don't keep your feet, there is no knowing where you may be swept off to." Bilbo Baggins ...

                Comment


                  #9
                  Re: Symptom question

                  Originally posted by kc8adu
                  should have put the real name with the common name after it.
                  btw the o.p is encouraged to post pics of the mess in that antec and the board.
                  Not sure if my digital camera is up to capturing the details but I'll see what I can do. I'll wait until the MB is out of the case though to get a picture. I'm thinking the top bulges will be much more apparent with an oblique shot.

                  Originally posted by PedroDaGr8
                  By the way the cap company is actually fuhjyyu and it tends to be the biggest problem. A simple quick visual inspection could give you all of the answers that you desire. No need to be afraid of it, just look in at the motherboard see if any of the caps are bulging or are of known bad companies such as GSC or OST etc., open the PSU (usually 4-6 screws and it opens fine) see for bulging caps. That will save you time and let you know what it could be. Rather than switching parts and waiting for an error to occur
                  The system is in an Antec case (at least the case hasn't given me any problems) so I have the side panel off as I'm typing. No problem checking to see what the capacitors on the MB look like. Checking the PSU will have to wait until I have a replacement since I'm off the net when the system is down (someday I'll look into having someone host me but then I don't have any more fun).

                  I'm just hoping the MB and PSU hold together long enough for me to make a decision as to how I want to replace them. Lots of options. I'm thinking I'll pick up a decent replacement PSU tomorrow and do a quick swap. If the PSU isn't the problem, I end up with a spare PSU. If the PSU is part of the problem, tomorrow is trash day.

                  Cheers,
                  Dave
                  Quitely implementing RFC 1925 wherever I go.

                  Comment


                    #10
                    Re: Symptom question

                    well it not going to fail over night
                    (although stranger things have happened)
                    But yeah get another "decent quality" PSU and try it since you basically want to keep the box in service
                    since you are in some respects abusing it it needs to be a decent one with decent caps but yeah server grades arn't cheap I suppose

                    I am not really the best guy to recommend a psu thought

                    But maybe a Seasonic s2 series or whatever it is consumer job I think the range all use decent jap caps

                    Corsair HX 620W is built on the seasonic guts not sure but think the HX520 uses jap caps too and seasonic built
                    VX series are from another psu builder but I believe still are good supply but may not quite be the same as hx in regards to caps
                    (I mention these cause they are a little cheaper but still somewhat a quality supply)

                    You requirement is "long term reliability" not grunt power of psu
                    (but a little extra never hurts)

                    Sounds like you going for the rebuild so may as well go for one you can reuse
                    Depending on what level of crap the other supply you mentioned you had is, it could tied you over (thats if you still got it)
                    post what it is for comment

                    MB caps sounds like they aren't too bad
                    (but caps can fail and show no physical signs of failure)

                    psu is more likely your trouble spot at this point
                    MB sound like its on the start of the slippery slope
                    (like KC8 said its more a preemptive move a recap now for the MB)

                    What you tend to find is as they start to fail you get more system hangs and reboots
                    it could take it a year (or more depending) to get to that level were its hanging rebooting very frequently.
                    (but as they go down the possibly becomes more a good chance of happening and usually right when you don't want it to)

                    So you probably got a little bit of breathing space with the MB.

                    Anyway a new psu will soon tell and make sure you crack open the old one to confirm the condition of caps.

                    keep us posted on your findings, the more conclusive outcomes we have the better

                    Cheers
                    You step into the Road, and if you don't keep your feet, there is no knowing where you may be swept off to." Bilbo Baggins ...

                    Comment


                      #11
                      Re: Symptom question

                      Most actual PSU`s are not intended for an high 5v & 3.3v load. So may be it is the smartest option to recap & refan the Antec.

                      Comment


                        #12
                        Re: Symptom question

                        Played vulture at the local CompUSA going out of business sale and picked up a CoolerMaster 500w PSU for $60. I figure I'll swap PSUs tonight. Since I'm not exactly an electronics handyman, I figure I'll swap out the MB and then see if I can fix it when breaking it doesn't mean that I'm off the 'net. Same for the old PSU. If I'm successful, I can retire my old "crash and burn" test system: a 733MHz PIII.

                        BTW, I'm wondering if the MB lasted as long as it did since the system is always on. I'm thinking the capacitors have had very few cycles of charging and discharging. That's about all I know about capacitors so they may have some other use on a MB.

                        Cheers,
                        Dave
                        Quitely implementing RFC 1925 wherever I go.

                        Comment


                          #13
                          Re: Symptom question

                          Originally posted by gonzo0815
                          Most actual PSU`s are not intended for an high 5v & 3.3v load. So may be it is the smartest option to recap & refan the Antec.
                          Not true with the older PSUs. That's only true with PSUs of 2005 and later, because of changed ATX specs. Because the later ATX specs use the +12V for the Vcore VRM.
                          ASRock B550 PG Velocita

                          Ryzen 9 "Vermeer" 5900X

                          32 GB G.Skill RipJaws V F4-3200C16D-32GVR

                          Sapphire Nitro+ Radeon RX 6750 XT

                          eVGA Supernova G3 750W

                          Western Digital Black SN850 1TB NVMe SSD

                          Alienware AW3423DWF OLED




                          "¡Me encanta "Me Encanta o Enlistarlo con Hilary Farr!" -Mí mismo

                          "There's nothing more unattractive than a chick smoking a cigarette" -Topcat

                          "Today's lesson in pissivity comes in the form of a ziplock baggie full of GPU extension brackets & hardware that for the last ~3 years have been on my bench, always in my way, getting moved around constantly....and yesterday I found myself in need of them....and the bastards are now nowhere to be found! Motherfracker!!" -Topcat

                          "did I see a chair fly? I think I did! Time for popcorn!" -ratdude747

                          Comment


                            #14
                            Re: Symptom question

                            I have a whole lot of KG7 in client service, and have recapped most of them.

                            470 16.0 8.0 20.0 4
                            1000 6.3 8.0 12.0 1
                            1000 10.0 8.0 15.0 15
                            1500 6.3 8.0 15.0 9
                            3300 6.3 10.0 25.0 6


                            The above table is from my spread sheet:

                            uF voltage diameter length count

                            It will look awful when the board compresses the columns. I recapped all the above with MCZ. IMO, it is worth recapping these boards, just to salvage the investment in XP-2000 processors.

                            Memtest86+ fails 100% on Test 5 for this board. Ignore it. It is a diagnostic bug peculiar to Athlons, AMD-761 and some other obscure conditions. All my boards operate propely, yet all fail Test #5.

                            Be sure to flash BIOS to the "DS" version.

                            Comment


                              #15
                              Re: Symptom question

                              humm yeah spec on psu not totally sure off hand "which is why the not the best guy thing"
                              and yeah spec etc have changed.

                              Cooler Master I have read of problem(s) with
                              (not sure of model(s) and don't recall what they were)
                              I nearly got one for this box (now using Corsair HX620W )
                              (it was the only thing in stock at the time and not all that cheap, cooler master)
                              If memory serves they don't have the best of caps (but few do),
                              I think problem(s) were more minor thing.

                              Still it should last a while and isn't a crap supply I don't think.
                              but again I'll stand to be corrected on the above re the Cooler Master PSU

                              Thing is, now you are aware of the issues and know what to look for if its starts getting flaky and you suspect HW prob

                              Now too is the test,
                              but like I said crack open that Antec to confirm stuffed caps.
                              that should help you rest a little easier about the system.

                              I would say the cycling (or lack of) also possibly really more so the heat remaining at a more constant rate
                              (not much expansion and contraction of the MB and components )
                              may have helped in its longer life and minimal issues
                              so keep it well ventilated and clear of dust.
                              (make,use dust filters of some kind if your not already)

                              When the VRM powers up,
                              the caps used for the CPU Voltage (output VRM caps) are changed via the VRM circuit at a limited rate.
                              (so inrush current to them is limited)

                              In other words their just not dumped across a voltage and allowed to suck as much current as they can.
                              Thats my understanding anyway

                              I should have mention , I have read of HDD stopping in server and never starting up again so hope you had everything backed up before shutting it down.
                              (only what I ve read but they are a mechanical device)
                              from the sounds of it you do regular backups anyway.
                              (again how much truth is in that I don't know but it sounds very possible to me)

                              Anyway like I said if you like the system and it works well for your needs
                              recap is the real best way to go for the MB
                              if it was me Id have a spare MB
                              (knowing me, probably system)
                              I could put into service

                              Maybe on you next computer fair hunt see if you can find a not too abused spare MB
                              (thats just a thought and of course you would have to make sure its works ..then get it recapped)

                              These days you can get a lot of grunt for not too much outlay so maybe thats really the way to go.

                              is this the MB?

                              http://www.techspot.com/reviews/hardware/abit_kg7raid/

                              http://www.ocworkbench.com/hardware/.../kg7raidp2.htm

                              just a thought really and only mention it cause of how you are using it.

                              The reason I asked was those chipset fans fail a lot from what I understand and although a non OCed system, the passive heatsink might handle it you might want to look a using a totally passive heatsink (little larger one) for the chipset.
                              (and thats if its got one)

                              Bgavin may have some input on this, and thanks mate for posting the cap list

                              like I said not the Guru here and these are only ideas that are open to correction or better ones

                              keep us posted

                              Cheers

                              Bgavin may have some input on this
                              Last edited by starfury1; 01-16-2008, 06:17 AM.
                              You step into the Road, and if you don't keep your feet, there is no knowing where you may be swept off to." Bilbo Baggins ...

                              Comment


                                #16
                                Re: Symptom question

                                Originally posted by starfury1
                                humm yeah spec on psu not totally sure off hand "which is why the not the best guy thing"
                                and yeah spec etc have changed.

                                Cooler Master I have read of problem(s) with
                                (not sure of model(s) and don't recall what they were)
                                I nearly got one for this box (now using Corsair HX620W )
                                (it was the only thing in stock at the time and not all that cheap, cooler master)
                                If memory serves they don't have the best of caps (but few do),
                                I think problem(s) were more minor thing.
                                I'm guessing that a lot of people got burned by bad caps in the supply chain. I seem to remember that CoolerMaster was one of the PSUs that Tom's Hardware liked at one point.

                                Originally posted by starfury1
                                Still it should last a while and isn't a crap supply I don't think.
                                but again I'll stand to be corrected on the above re the Cooler Master PSU

                                Thing is, now you are aware of the issues and know what to look for if its starts getting flaky and you suspect HW prob

                                Now too is the test,
                                but like I said crack open that Antec to confirm stuffed caps.
                                that should help you rest a little easier about the system.
                                May have to wait until this weekend unless a crash forces my hand. I've been job hunting for a while and I have a short notice interview tomorrow.
                                Originally posted by starfury1
                                I would say the cycling (or lack of) also possibly really more so the heat remaining at a more constant rate
                                (not much expansion and contraction of the MB and components )
                                may have helped in its longer life and minimal issues
                                so keep it well ventilated and clear of dust.
                                (make,use dust filters of some kind if your not already)
                                I should really move the server to the basement. My cat likes to hang out with me when I'm computing. I'd swear she sheds continuously.
                                Originally posted by starfury1
                                When the VRM powers up,
                                the caps used for the CPU Voltage (output VRM caps) are changed via the VRM circuit at a limited rate.
                                (so inrush current to them is limited)

                                In other words their just not dumped across a voltage and allowed to suck as much current as they can.
                                Thats my understanding anyway
                                I don't have the foggiest idea how stuff works at the electronics level on a system. Some day I'll probably take the time to learn what's going on at the trace level.
                                Originally posted by starfury1
                                I should have mention , I have read of HDD stopping in server and never starting up again so hope you had everything backed up before shutting it down.
                                (only what I ve read but they are a mechanical device)
                                from the sounds of it you do regular backups anyway.
                                (again how much truth is in that I don't know but it sounds very possible to me)
                                All of the critical drives are Linux software RAID 1 (mirrored). At least when I put the system together I could power down, pull a ribbon cable off of any drive and then power up and the system would work.
                                Originally posted by starfury1
                                Anyway like I said if you like the system and it works well for your needs recap is the real best way to go for the MB
                                if it was me Id have a spare MB
                                (knowing me, probably system)
                                I could put into service
                                Assuming I can successfully recap the KG7-RAID, that's what it will become.
                                Originally posted by starfury1
                                Maybe on you next computer fair hunt see if you can find a not too abused spare MB (thats just a thought and of course you would have to make sure its works ..then get it recapped)

                                These days you can get a lot of grunt for not too much outlay so maybe thats really the way to go.

                                is this the MB?

                                http://www.techspot.com/reviews/hardware/abit_kg7raid/

                                http://www.ocworkbench.com/hardware/.../kg7raidp2.htm
                                That's it. The Highpoint RAID is fairly worthless. Back to the test mentioned above, if I tried it with the Highpoint RAID instead of Linux software RAID, the Highpoint BIOS would detect the missing drive and then go to a screen that asked whether it should continue or not. No way to set it to continue by default or after some delay. So I used the Highpoint RAID as JBOD.
                                Originally posted by starfury1
                                just a thought really and only mention it cause of how you are using it.

                                The reason I asked was those chipset fans fail a lot from what I understand and although a non OCed system, the passive heatsink might handle it you might want to look a using a totally passive heatsink (little larger one) for the chipset.
                                (and thats if its got one)

                                Bgavin may have some input on this, and thanks mate for posting the cap list

                                like I said not the Guru here and these are only ideas that are open to correction or better ones

                                keep us posted

                                Cheers

                                Bgavin may have some input on this
                                Picked up a new MB (MSI K9N), CPU (Athlon 2X 5600+), RAM and video. I'm thinking I'll swap my old desktop (Tyan Tiger MPX with 2x Athlon 2400+ MPs) into the server role and use the new MB, etc. for my desktop. That minimizes my down time since I can pretty much just pull the drives out of the old server and stick them into the Tiger enclosure. I don't really need an SMP server since the old 1700+ could keep up but that's the easiest upgrade path. I'll play with the KG7 board "off line" when putzing with it for several days or weeks isn't a problem.

                                Cheers,
                                Dave
                                Quitely implementing RFC 1925 wherever I go.

                                Comment


                                  #17
                                  Re: Symptom question

                                  As to:

                                  >>> I should really move the server to the basement. My cat likes to hang out with me when I'm computing. I'd swear she sheds continuously <<<

                                  Passing thought.
                                  Shave the pussy for better cooling. ~ ~ ~ Ahem....

                                  .
                                  Mann-Made Global Warming.
                                  - We should be more concerned about the Intellectual Climate.

                                  -
                                  Be who you are and say what you feel, because those who mind don't matter and those who matter don't mind.

                                  - Dr Seuss
                                  -
                                  You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
                                  -

                                  Comment


                                    #18
                                    Re: Symptom question

                                    Originally posted by PCBONEZ
                                    As to:

                                    >>> I should really move the server to the basement. My cat likes to hang out with me when I'm computing. I'd swear she sheds continuously <<<

                                    Passing thought.
                                    Shave the pussy for better cooling. ~ ~ ~ Ahem....

                                    .
                                    Not a good idea. She still has her claws and knows how to use them:
                                    Quitely implementing RFC 1925 wherever I go.

                                    Comment


                                      #19
                                      Re: Symptom question

                                      Ouch!
                                      Mann-Made Global Warming.
                                      - We should be more concerned about the Intellectual Climate.

                                      -
                                      Be who you are and say what you feel, because those who mind don't matter and those who matter don't mind.

                                      - Dr Seuss
                                      -
                                      You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
                                      -

                                      Comment


                                        #20
                                        Re: Symptom question

                                        I run Alpha PAL8045 coolers on all my KG7 boards. This is the massive HSF that bolts through the board using the four holes around the CPU socket. I use Panaflo 12U series fans, and they never let me down. The temperature delta averages 11C on most of my machines.

                                        I've been installing Sparkle ATX-350PN supplies for these boards without a problem. This is the PSU with the 120mm fan. It is populated with TEAPO and OST caps in the larger capacities. I have not yet found a source for Rubycon ZL, so I have not bothered recapping any of the Sparkle yet. They run quite cool.

                                        I run a customized version of the DS bios on these machines. Mine sets the Safe and Performance mode defaults to values that work for me. When this board first came out, the advice was to change the SLEW from Auto to 7 (Max). I've been running all these boards this way for years, and they are all stable.
                                        Last edited by bgavin; 01-17-2008, 06:38 AM. Reason: edited for typos

                                        Comment

                                        Working...
                                        X