PeterW (14th March 2009)
Further from a post I won't hijack any further, thought I'd run this question past yourselves:
I mentioned a site with around 30 of your units with rather large reliability issues.
I spent a good deal of the day on Thursday looking into this properly and this is what I've found.
Many of them (over half so far are definitely confirmed) freeze entirely during the Windows load process. Restarting will either see them crash at POST (eliminating a software installation issue) or continue onto windows and freeze again at around the same point, randomly. The only way around this is to unplug the unit from the wall (or flick it off at the PSU switch) for 30 or so seconds. Then, they'll usually boot Windows without problem. By this time however, due to unsafe shutdowns, they exhibit failures and problems such as desynchronisation with the GPO and what not, but that's software related (although a direct result of the above HW issue so won't go into it further)
So, I didn't want to phone up SCL and have a shout (I won't anyway, having a fairly bad speech issue rules out using phones ) without running this past you chaps first as some light may be sheddable.
There are two possible things that are crossing my mind: 1. rough electric supply. This site is out in the sticks, and the UPS on the server is often trimming irregularities with said supply. I've thrown a multimeter on it for half an hour and logged the results, they do waver as expected with AC but within limits, bearing in mind the short testing timeframe (and me not being an electrician)
And 2. a faulty/crap batch of power supplies. PSUs are supposed to compensate for normal AC irregularities to a small degree, some do it better than others and some don't bother at all (hello Q-tec!)
So, is there anything you can suggest I look at on my next visit to either remedy or properly diagnose where the fault may lie? I'm tempted to grab a PSU tester and run one of the unit's PSU's on it's own for a good couple of hours to get a better picture of that (alongside a Seasonic or something for comparison), but don't really want to go popping open units under warranty without good reason.
Any suggestions gratefully received in which case, I don't really want to have to either RMA the entire suite leaving them with nothing, nor leave it as it stands and have to spend all my SLA'd visits rebuilding machines. The ICT coordinator did suggest she flipped the trip for the suite every night however that's likely to cause more issues further down the line - time to buy shares in CR2032 batteries for the bios
Of course suggestions welcome from others outside of SCL too as this may have been seen elsewhere with other hardware. You might have guessed from previous posts that I'm really, really not a fan of MSI which never helps
Well explained and outlined post, so nice one there.
Out of curiousity, have you tried taking one off site and building it which would entirely eliminate the question of poor power supply within the site; not likely but it helps to go that extra mile.
Pretty dire performance there you've got; have you tried a burn in CD to see whether you can get the unit to fail with a decent indicator of the issue? I find it really handy to give companies a screenshot of the failure on a burn in tester so as to help strengthen the case for rectification. - Mostly memory and SMART errors though.
I have no doubt SCL will pick this up and sort it sharpish.
Not taken one off site as yet, may well be the next step. Saying that, rather than leave them without a machine for a fortnight I might be inclined to borrow a UPS to take over, or hook it into the switch UPS so there's an extra bit of protection and filtering involved.
Not run a stress test as yet either: I did leave a memtest running on loop for 3 hours but that came up fine, but a memory failure was only in my mind when it was 1 machine doing it. I've got some all round stresstests but they're all windows based - I might have a look on my bootable discs and see if I've something else (ala PC Check) that'll warm up it sufficiently to cause a repeatable problem.
I know we had HUGE issues getting these things to build on CC3 in the first place, I'm fairly sure the office won't be happy to see one come back in Odd really, as the mainboard in place is next to identical to the mainboards in RM One kit short of a BIOS revision (something else that crossed my mind was a bios update but really didn't want to start fiddling with that unless really needed or suggested by SCL/MSI)
Cheers for the continued input btw.
If you email Peter at the company in question and explain as above, and mention the speech issue, I am sure he will sort it pretty quickly.
Not sure about the fairness of mentioning the company name as there is a chance that it is nothing to do with the machines and/or something that sometimes happens like a bad batch of psus - but others thinking of using the company may presume that they sell dodgy kit and avoid them when actually they are a very good company and would no doubt love a chance to sort your problem out.
You might consider removing the company name whilst you get some feedback from members to be fair.
Cheers for the input but as I'd already posted in another thread (which you may not have seen) I'd gone off on one and called all of their hardware unreliable rubbish beforehand (edited later on) so mentioning a name now won't be an issue, unless SCL specifically want me to
Either way, bad batches/faulty hardware isn't indicative of a companies performance or reputation and never has been. How they deal with it is, and I've little doubt of SCL's competence there.
And cheers, yeah I might well email drop Peter an email, will have a hunt for some more details soon. Bit of a pain only being onsite for half a day a fortnight :/
Let me first apologise for the isssues you appear to be having with a set of SCL computers we certainly don't like to hear that your time has been taken up trying to correct hardware isssues which you shouldn't be having in the first place.
I'm about 90% sure I know which site you are reffering to due to this(on a side note I may have a solution for this which we have been testing)I know we had HUGE issues getting these things to build on CC3 in the first place
Please if you could PM or email me email@example.com me the site name or a PC serial number I can get all the details to hand, I wouldn't recommend updating any boards BIOS unless you are 110% sure you know which one you need as some of the boards we use have multiple versions which all have thier own individual BIOS versions. The issue does sound strange and the number also sounds very high, from the top of my head I don't think the hardware is anything out of the ordinary so I would have expected to hear from other customers if there was a widespread issue.
If possible I would like to have one of the effected machines collected asap so that we may have a proper look at it and get to the bottom of the issue.
I am aware you personally have not long been looking after this site for very long but I would just like to note this issue has never been reported to us but I am glad you have brought it to our attension.
Don't be a stranger please get in touch.
10 out of 10 for that Peter, that's what I like to see.
Not back on site till a fortnights time, however I've set up a couple of "tests" for the ICT co-ordinator to keep a watchful eye on in the meantime as well as asking them to organise for an electrician to check the power supply in that room specifically (an independent sparky to those who fitted the supply naturally). Being the only common denominator likely to cause such an issue I want to at least eliminate the electrics as the cause before going through god knows how many machines/causing more work for myself and SCL + problems for the school.
I might drop by the school later this week as I suspect I'm just up the road for something else to check up on them. If I do, I'll grab some details from them such as serial numbers and what not, but I'd really rather get the electrics checked first before we start wasting someone's time unnecessarily
As soon as I get the info I'll drop you an email.
Cheers for the support by the way, it's good to have a place such as this to get heads together, especially when you're really unsure where the fault lies
You don't really need to search for a possible reason for your problem because the company in question will do it for you, and IMHO you really should be speaking to them directly
Last edited by mark; 15th March 2009 at 12:22 PM. Reason: superfluous post removed
Man these Edumods are a serious bunch.
I had to look up synaesthesia and it s good word... so is pretentious.
We all have bad days, hell, I reckon most have slagged companies off and its the way they react that shows how good they are.
I have similar issues, but on a smaller scale, SCL have posted spare out when asked without question.
Didn't mean to provoke such reaction!
I've already made a very public apology. I've already stated the reason why I said why I did - thinking before typing, an problem I often have to keep under control and occasionally fail.
If anyone can draw from this thread the conclusion that SCL is doing anything wrong, then I shall request from a moderator it's closure. And why am I asking on here rather than just going to SCL straight? Because, if you'd taken the time to read the thread in entirety, there's a strong possibility that this isn't down to their hardware and I'm trying to eliminate all possibilities before taking up their time unnecessarily.
So, "thanks" for referring to me as a "cad". Not everyone is the same and on some days I go through life preferring to have a terminal illness than the mental problems I have. Most days however I have the courage to call myself a ruddy moron and grow up.
Regardless of my problems, SCL are doing a damned fine job already and I've already PM'd Peter with specifics. I have little doubt that, should the problem be down to their hardware, it will be sorted out. Bad things happen to all companies, everyone gets negative comments now and then, and if noone had any negative feedback they would never have scope to improve. However, it's the way of things that you always focus and only hear about the bad things. Google ANY company name and you will find a few hits saying "don't go near them!" "They're dreadful!" etc etc. You'll rarely come across anything that says what a good experience someone has had, how out of their way someone has gone to help with a problem.
Take Seagate for example. Hugely respectable brand, and they've just suffered a large batch problem with some of their large capacity hard drives failing. The internet fills with "Seagate are bad, don't buy their stuff ever again! propoganda before you can blink, and that sort of rubbish fills the forums before you can find "actually, they've gone out of their way to offer free replacement AND several thousands of pounds worth of free data recovery to anyone effected"
So, rather than venting off, try giving a thumbs up to SCL and PeterW. Feel free to berate me for a momentary collapse of composure. I'm only human.
Oh, and ICT Guy - that's an ungodly hour of the morning!
There are currently 1 users browsing this thread. (0 members and 1 guests)