AFH
May 4, 2019 15:14
I automatically set swappiness to 10 without thinking about it. Reducing swap will have a similar effect, but once the swap is full things grind to a halt. Without knowing what is being run, it's impossible to say whether 8GB RAM is enough. A VM or two, an IDE and some Wine programs will soon fill that.
AFH
May 4, 2019 15:14
@davidgo - We'll have to differ on the effect of increasing the swap space: I cannot see how it can possibly do harm, and it's very easy to test the effect of increasing it. And the questioner's description of his freeze is exactly what I have observed when swap has run out: the code cannot get into to memory even to allow processes to be killed. By the way, the rule of thumb that the swap size should be double the RAM size is nonsense: the less RAM you have, the more swap you need. The swap and RAM size combined must be plenty for the maximum number and size of active processes.
AFH
May 4, 2019 15:14
That's not a lot. Running out of swap space is a frequent cause of freezing. I suggest you try a much larger swap, eg 16GB. You don't need to re-partition: you can use a swap file instead, the only disadvantage being that you won't be able to hibernate.
AFH
May 4, 2019 15:14
How much swap have you configured?
 
AFH
Jun 13, 2018 22:07
@robinCTS As for prefacing my comments, you are correct, but after testing with LO I meant to confirm with Excel: I just forgot.
AFH
Jun 13, 2018 22:01
@robinCTS Done.
AFH
Jun 13, 2018 20:25
I leave it to you to decide which comments are worth leaving in.
AFH
Jun 13, 2018 20:21
I did my basic testing in LO, but confirmed in Excel 2007, the most recent version I have access to. Unfortunately, I was so pleased to have found that solution that I forgot to cross-verify it. Sorry.
I was using ordinary cells, not cells in a Table: there's nothing quite like that in LO5, but it is under consideration for LO6, which is why I didn't quite follow you at first when you referred to a Table. Thanks, I've learned a lot.
AFH
Jun 13, 2018 16:51
I still haven't found a circumstance when ROW($1,$1) returns anything other than 1 or any cell where it gets incremented/decremented when a row is inserted/deleted.
AFH
Jun 13, 2018 16:46
My use of MATCH() was more to prove I could do it than to provide a usable answer: to be honest, my original suggestion of =IFERROR(SMALL(A:A,ROW()-1),"") seems the simplest, and the use of IFERROR() means that the column can be pre-filled beyond the current number of rows in the data.
AFH
Jun 13, 2018 16:45
There may be better ways of finding the first number in a column. I use regular expressions quite a lot, so their use comes fairly naturally to me, but I'm the first to admit that they look geeky, and beyond the needs of the average Excel user. I use a lot of the Excel functions, but I've never explored more advanced data structures and features, and I don't use VBA, since I mostly use LibreOffice on Linux.
AFH
Jun 13, 2018 13:58
I've found a formula that seems to work: =SMALL(A:A,ROW()-MATCH("^-*[0-9][0-9]*$",A:A,0)+1). The MATCH() finds the first numeric value in the data column and uses this to offset the ROW() in the sort column.
AFH
Jun 13, 2018 13:58
I've tried this, but ROW($1:$1) still returns 1 when an extra header line is added. What sort of works is =SMALL(A:A,ROW(1:1)), but only the row below the inserted line(s) is maintained with ROW(1:1). The rows below are unchanged, so the formula has to be copied again down the rest of the column. If a row is inserted above the existing header, none of the formula cells is maintained. What works a bit better is to set a work cell (I chose G1) to =ROW(A2)-1 and use =SMALL(A:A,ROW()-G$1) for the sorted column, but even this fails if a row is inserted above the header (G1 moves).
AFH
Jun 13, 2018 13:58
Nice answer, but I found that =IFERROR(SMALL(A:A,ROW()-1),"") works equally well. What's the advantage of adding MATCH() and INDEX()? And ROW($1:$1) always returns 1, so why not use the literal, which can be incremented if there are more header lines?
 
AFH
May 4, 2018 20:07
The closest you'll come is to use dual boot and hibernate each system when switching, but you won't be able to see the discs mounted to the other OS, and there will be a delay during the switch. A VM will give you a faster switch, but with the performance costs you have already noted.
 
AFH
Dec 20, 2017 07:37
@Aternus - As I said you need only kill Explorer (quickest, but with side-effects), though log off and on will be more orderly and quicker than a full restart, but all processes running in the session will terminate and need to restart.
AFH
Dec 20, 2017 07:37
I have been having random crashes of Explorer ever since XP, but as I said, I haven't noticed that much difference with 1709, though these days my day-to-day OS is Ubuntu, so I am now using Win10 and Explorer a lot less than I used to.
AFH
Dec 20, 2017 07:37
You have done the sensible tests, but it looks like an OS bug, so you'll need to wait for an update. It's something I do reluctantly, but if you kill explorer.exe it will restart and the OS will continue running, although your desktop will be messed up, especially the system tray (missing and reordered icons). Sometimes Explorer crashes without my having to do it manually, and it nearly always gives an error on shutting down during a restart. I have learned to live with it, though I have not noticed any significant change since 1709.
 
AFH
May 28, 2017 07:59
OK. I didn't realise that the symptoms had gone away permanently.
AFH
May 28, 2017 07:59
One thing to try is booting an independent OS, such as a Windows diagnostic disc or a Linux Live CD. If there is a hardware problem, you should see the same symptoms. Also check if you can disable the camera in your BIOS: laptop cameras are usually interfaced though an internal USB hub.
 
AFH
May 26, 2017 08:50
Once you have the copy, you can restore it to a fresh drive and then whatever files are recoverable should be on it. Make sure you write-protect the copy, as you may never get another one.
AFH
May 26, 2017 08:50
I have read in several places that a failing drive can be resuscitated temporarily by putting in a freezer for several hours (after wrapping it in several layers of moisture-proof polythene). The fact that your drive came back to life briefly suggests that this may be worth trying. Before you remove it from the freezer and connect it, be prepared to copy the drive: you can use something like sudo dd if=/dev/sdX of=TargetFileOrDevice bs=64M conv=sync. The last clause is important, to prevent disc errors offsetting the data on the copy, but it may make the output file slightly too large.
AFH
May 26, 2017 08:50
It wasn't the OS crash that caused the damage; rather, the disc suffered a major hardware failure and this caused the OS to crash. The cause of the failure could be overheating or another cause of failure of one of the mechanical parts: motor, bearings, head servo, dust intrusion, etc. It is probably worthwhile to let one of the other data recovery firms have a look at it: if the problem is simpler than your first analysis suggests, then they ought to give you a much lower cost for recovery. Otherwise, follow @Tetsujin's advice.
 
AFH
Apr 17, 2017 20:27
Yes, you're right. The TCC-LE solution (start /elevate) has the same effect, as does my solution. The privileges appear to be attached to the window, rather than the process. So you're quite right when you say it can't be done in the same window, but if the program is native Windows a new window will be needed anyway.
AFH
Apr 17, 2017 20:27
I don't believe the questioner wants to change the privileges of any running program, but instead wants to start an entirely new process with elevated privileges.
 
AFH
Jun 21, 2016 11:39
I notice that there have been a couple of up-votes for your question, so there is some degree of interest in it. Since we have reached a conclusion of sorts, I think it ought to be documented, so I propose to post an answer, but I'll leave it for some hours, in case you have anything else to add.
AFH
Jun 21, 2016 01:32
I'm going to bed, now. If I can help further, please contact me and I'll try answer you tomorrow.
AFH
Jun 21, 2016 01:24
Yes, I did see it. I guess you have the work-round, though I'd be careful: if they realise what you're doing they may cut you off for breaching of your terms and conditions.
AFH
Jun 21, 2016 01:18
Port 1000 was fine, but not port 80. Maybe you're right about your ISP - maybe they don't like customers running web servers.
AFH
Jun 21, 2016 00:56
Can you check the logs on the web server to see if it receives any contact from outside? Maybe check the router logs as well, and compare the failed web access with the successful FTP access.
AFH
Jun 21, 2016 00:22
And the iPad works on the internal network?
AFH
Jun 21, 2016 00:22
Yes, precisely.
AFH
Jun 21, 2016 00:22
Have you tried mapping a different port, as in FTP and SSH (at least, I think that's what the line above is)?
AFH
Jun 21, 2016 00:22
You have other services set up on the same IP: do they work? Are there any firewalls involved?
 
AFH
Jan 11, 2016 16:14
I am not sure I gave you an answer, only suggestions of things to try. You know best the details of what worked for you, so it's probably better to submit the answer yourself (an acknowledgment would be nice). Glad to have been of help.
AFH
Jan 11, 2016 16:03
OK. Is our discussion now at an end? I can't see that I have much more to contribute.
AFH
Jan 11, 2016 15:58
OK. I have been around long enough to avoid the first release of any new product: first customers most often turn out to be gamma testers (not just computers, I use the same criterion for everything, from cars to televisions to washing machines).
AFH
Jan 11, 2016 15:35
You don't need to wait for GF1000: there are plenty of other cards currently available which support Direct3D, some quite inexpensive. Since you're replacing integrated graphics, performance should not be a primary concern.
AFH
Jan 10, 2016 17:11
As for Direct3D, you're well out of my experience, but I did find several links which indicate that you will need a new graphics card to replace the integrated graphics, which is probably impossible if you have a laptop: see Asmodai's thread here.
AFH
Jan 10, 2016 17:10
OK, so it's not the licence that expired, but your right to use it. Didn't your machine come with a Windows licence? You should be able to use this for upgrading, unless it's pre-Win7.
AFH
Jan 9, 2016 00:43
I didn't know that MS issues licence keys that expire.
AFH
Jan 9, 2016 00:17
I'm not suggesting re-partitioning, but changing the flag on an existing partition (right-click on the partition, then select Manage Flags.
AFH
Jan 8, 2016 22:51
You can take a copy of just that partition and restore it if the upgrade works. But try gparted first.
AFH
Jan 8, 2016 22:42
Have you tried bootrec /fixboot? Alternatively, since the tag indicates that your Linux system is Debian, you should have gparted available, and this is able to set the partition flags. The EFI partition has both boot and esp set, and most of the others (including recovery) have msftdata set only, except that the Linux swap partition has no flags. I am not sure how these flags relate to the diskpart report.
AFH
Jan 8, 2016 18:17
This link seems to deal with UEFI repair for all disc types. I know you can boot, so your boot loader doesn't need repairing as such, but these instructions may get it into a better state for upgrading to Win10.
AFH
Jan 8, 2016 01:05
The link is here. Although it refers to Windows 7, it may be relevant to 8.1.
AFH
Jan 8, 2016 00:55
I found a link which suggested that getting rid of grub can cure installations problem, but your question implies that you are not using it.
AFH
Jan 8, 2016 00:43
2016-01-07 23:40:52, Info CONX ConX::Compatibility::CSystemAbstraction::HostIsUEFIFirmware: Host is UEFI.
2016-01-07 23:40:52, Error CONX ConX::Compatibility::CSystemAbstraction::HostIsUEFIDiskLayout: Failed to retrieve system partition NT path.
AFH
Jan 8, 2016 00:43
The following two lines from the middle of the log seem to to show when things start to go wrong.
AFH
Jan 7, 2016 10:31
A passing thought: how is your disc partitioned? Is it possible that you have a small C: system disc and a large D: data disc? If there were insufficient space on C: for the new system and back-up files, an upgrade in situ would be disallowed.