| InHisName | Oh so quiet here tonight | 02:07 |
|---|---|---|
| ChinnoDog | @crickets | 03:18 |
| InHisName | That memory has some pretty strange corruptions. May have to break down and spend some unemployed $$ to replace it. Sniff sniff. | 03:28 |
| ChinnoDog | I thought we concluded your motherboard has bad address lines | 04:08 |
| waltman | So at a family gathering yesterday, some relatives were talking about how they'd filled up their android phones' memory with too many emails. | 04:12 |
| waltman | My brother described his complex system where he deletes some mail on his phone but not on his backend mailbox. | 04:13 |
| waltman | But on my iPhone, I just have it set to keep the last 50 messages on each mailbox. That works fine. Since the backend is imap, if I ever need to pull up something older I can just do a search. | 04:14 |
| waltman | Do android mail apps work differently? | 04:14 |
| waltman | Or is $brother just demonstrating that he doesn't understand how things work? | 04:15 |
| waltman | I should point out that when topics like this come up, my defense mechanism is to pretend that I don't know anything about computers. :) | 04:21 |
| InHisName | No, ChinnoDog, the problem followed the memory stick when moved to other slot. NOT motherboard. | 04:33 |
| InHisName | Oddly, I just noticed that the two memories are NOT the same. One is1GB PC2100 DDR CL2.5 266MHz and other (the one giving me fits lately) is 1GB PC3200 DDR CL3 400MHz. I looked up MB specs and first is its needs. | 04:37 |
| ChinnoDog | waltman: yes, it is that simple. Good tactic, I dont play dumb but I keep my mouth shut about these things | 04:50 |
| JonathanD | Morning. | 10:30 |
| rmg51 | Morning | 10:37 |
| waltman | Morning. | 12:50 |
| teddy-dbear | Morning peoples | 13:10 |
| jedijf | waltman: your strategy is best - supporting people is terrible | 13:29 |
| InHisName | Morning all | 13:49 |
| jedijf | morning wrong ram | 13:49 |
| InHisName | I have 3 partitions: / , /boot, and /home - I need a HOWTO link on reformatting / BUT not touching a hair on /home during the install from scratch etc. | 13:51 |
| ChinnoDog | iirc a reinstall does not reformat /home even though it tells you it is going to delete everything | 14:20 |
| ChinnoDog | Don't rely on me for that info though. I haven't done it like that yet. | 14:21 |
| ChinnoDog | I should have said that differently. If you select your /home partition for /home in the partition layout editor and do not tell it to format it you will get a warning that all directories will be deleted if you continue. I don't think that applies to /home though. | 14:23 |
| InHisName | I found a link: http://ubuntuforums.org/showthread.php?t=2011400 | 14:30 |
| jedijf | rsync home somewhere else first - safe sex - hope is not a method | 14:32 |
| InHisName | One time a user asked about preserving the programs installed by apt-get. It was not answered. Just the same answer as /home preservation was given to answer the original question. | 14:32 |
| jedijf | that was ChinnoDog, and he answered and resolved himself | 14:33 |
| jedijf | at some point everyone needs to do that ^^^ | 14:33 |
| InHisName | I'm not sure preserving the installs is even possible. Might be easier to produce a list of what was installed and keep that and reinstall it all after ubuntu fully built up. | 14:33 |
| jedijf | nice | 14:33 |
| jedijf | that's how i would approach it | 14:33 |
| jedijf | dpkg is your friend | 14:34 |
| ChinnoDog | The problem is the packages that don't follow the rules. In theory everything is heirarchical and you can cleanly add and remove whatever but in actuality some packages don't get removed automatically and don't go backwards easily. The result is that adding and removing packages on two different machines in a different order leaves you with a different list of installed packages. | 14:35 |
| InHisName | Old issue review: if a motherboard likes 266mhz memory and I had one 266mhz and another 400mhz, how is it possible that they worked for several years ? Is it possible that the 400mhz memory that is acting goofy, is really still ok? | 14:35 |
| ChinnoDog | There is no issue putting faster memory into slower slots | 14:36 |
| InHisName | so a working 400 in a 266 system is perfectly fine ? | 14:37 |
| ChinnoDog | Yup | 14:37 |
| InHisName | So when a 400 in a 266 is acting goofy, then it is likely just plain bad and not a fish out of water ? | 14:37 |
| ChinnoDog | That is correct | 14:38 |
| jedijf | the old downward compatable - it will be as fast as the slowest ram | 14:38 |
| jedijf | lcd | 14:38 |
| jedijf | in all of this did you test them individually | 14:39 |
| InHisName | Price of 'low-density' 400mhz 1GB memory was just below $20 / stick. The 266 1GB one was below $10. I ordered one. In a few days I'll have my backup pc back to 2GB. | 14:39 |
| jedijf | is that max? | 14:40 |
| InHisName | I memtest86+ each memory in each slot. 400mhz stick always failed regardless of which slot. | 14:40 |
| jedijf | i need some ram too for a newly acquired old laptop | 14:40 |
| jedijf | to take to 4 gig | 14:40 |
| InHisName | 2GB is max to stuff in that MB. | 14:40 |
| InHisName | google is your friend | 14:41 |
| InHisName | or ebay search | 14:41 |
| jedijf | yeah, usually just pricewatch and go with something average priced | 14:41 |
| jedijf | not a ram fan | 14:41 |
| InHisName | but ebay is lower on 'obsolete' pulls. Pricewatch good for NEW memory. Or old stock new memory. | 14:42 |
| jedijf | hmm...it *is* cyber monday | 14:42 |
| InHisName | 2 days ago I saw an ad for 27" montitor 2500x1900 pixels or so. Better'n HD tv specs. for $199. Need a job first. | 14:43 |
| InHisName | Been shopping for ipad retina (gen 3 or 4) used. Horrors! but see lots of notes that many apps for disabled and autistic children and really catching on. | 14:46 |
| InHisName | Know any really cheap sources / used seller I can check into ? PS no 'fencing' sellers if I can tell. Hard to discern on craiglist. | 14:47 |
| ChinnoDog | You can stuff more than the approved amount of memory in most motherboards. So long as the chip density is not to high it will work. | 14:54 |
| ChinnoDog | I do it all the time. I have 8gb of memory in my ThinkPad T400 I use for work and 4gb in my old HP laptop that is only approved for 2gb. | 14:55 |
| InHisName | Does it see more than 2GB? | 14:59 |
| InHisName | My stick is listed as 64x8 dual rank - so there is 16 chips. But I was only finding 128x4 - the site lists a jillion mb that they do NOT work with. Probably the workable boards are much smaller list. They used terms such as 'high density' Finding the 'low density' was tougher as many were not listed that way. | 15:05 |
| === stump_ is now known as stump | ||
| InHisName | This memory has normal fail from 1072 to 1111 with a stuck bit 0x00008000 every 65535 Start 0x00042uw1cc4 and initially ended at 1088 ( 0x00043yz1cc4). So I marked out that 32Meg. Still seemed to be flakey. Tested further. | 15:14 |
| InHisName | Wierd stuff: starting at 1428 (0x00059400000 ) I got errors ONLY on test #7 random. The measured value was a walking '0' thru ffffffffff. Strange for random data being written. It would only fail if I started test #7 at least 40 Megs before the failing start point. Otherwise it would pass. Not even sure if this is a real failure or what. So goofy and funky, that replacement is easier dependable route. | 15:18 |
| InHisName | ChinnoDog: any thoughts on why its behaving wierdly ? | 15:18 |
| ChinnoDog | Yes, it sees all of it | 15:30 |
| ChinnoDog | You can usually tell if the density is right by counting the chips on the memory module | 15:34 |
| InHisName | So a stick with 16 chips both sides, 1GB+ 266Mhz+ will probably work ? I ordered a 16chip 1GB 266mhz one for $9. It might make it here by sat. Sent from CA. | 15:40 |
| ChinnoDog | Density is what matters. If there is an approved module of the largest size supported by your motherboard with 8 chips on it then a module double that size with 16 chips on it will probably work. | 15:45 |
| InHisName | ChinnoDog: if I have starting address 0x00043001cc4 and ending address 0x00045671cc4. Too setup excluded mem for grub2... I start with 0x000430000000 and use a mask to run up to 0x00045fffffff mask: 0x0004?0000000 I have to pick value to mask for 3, 4 , and 5 or do I need to mask at 2**n boundry ? Then I reverse the bit pairs within the HEX for that big/little indian thing. So if it was 0111 then --> 1101 | 15:46 |
| ChinnoDog | A module double the maximum size with only 8 chips on it is double the density of the comparison module and probably won't work. | 15:46 |
| InHisName | Some MB are limited to single/dual ranks (chips both/ one side) I have seen 32 chips on a stick 16 on a side, but that is REALLY OLD memory. | 15:48 |
| InHisName | The 1GB has 'highdensity' 128x4 and mine is "low" density is 64x8 dual or 16 total from both sides. So if I picked up a 128x8 dual rank for 16 total it would still work but claim to be 2GB stick ? And I'd only see 1GB. | 15:51 |
| InHisName | BUT all of the higher sticks cost more $$. Of course applies well with free gifts and junk box discoveries. | 15:51 |
| InHisName | For the masking, I suspect I have to do it in three parts with 3 starting addresses and 3 masks. Because it does not start on a 2**n boundry nor end on one. | 15:59 |
| ChinnoDog | "high" and "low" density is relative to the time you are buying the chip since chips are constantly increasing in density | 16:16 |
| === LordOfTime is now known as TheLordOfTime | ||
| InHisName | I just found out that the memtest86 installed by 12.10 has bugs in it. v4.20. Makes test #7 fail wierdly and at every address. | 18:09 |
| InHisName | Running the memtest86 on the live CD for 12.04, is the same version v4.20 and works much better. No more weird errors. Now to find where the real errors are temorarily until the new stick arrrives. | 18:10 |
| ChinnoDog | That is terrible. memtest86 is not supposed to have bugs | 18:11 |
| InHisName | Well it's in launchpad, workaround is use older live CD. Strange thing is they did it right with 12.04 and it is with the same version (v4.20) used by 12.04 & 12.10. Go figure, huh? | 18:16 |
| InHisName | When I googled on the test #7 to find more to understand what was going on, I bumped into many reports of bugs with 12.10's memtest86. | 18:18 |
| InHisName | Now range is from 1056Meg to 1111.7Meg I think that is a total of 40Meg to toss away. I can do it in 3 starts & masks. Done 3 passes, will let it do at least 2 more before I set it up for good. | 18:22 |
| ChinnoDog | When I excluded memory on my system I blocked out the entire last chips. (Chips is plural because it was running in dual channel.) | 18:28 |
| InHisName | Now that I've tested a bunch of addresses, I can't get the bad section to fail anymore. | 19:07 |
| InHisName | I couldn't have fixed it could I ? Wait 10 min to 'rest' and try testing bad area again. | 19:08 |
| ChinnoDog | It might be intermittent | 19:21 |
| ChinnoDog | imho you should block out the entire range of the bad chips and see if it continues exibiting problems | 19:21 |
| InHisName | right now just running memtest86+, but will adjust the block soon. | 19:42 |
| InHisName | The block only applies to booting ubuntu from the hard drive. memtest86 seems to be a separate partition on the live CD, so it does not have the block setup. | 19:44 |
| === LordOfTime is now known as TheLordOfTime | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!