In message <u323if$27gt1$1@dont-email.me> at Thu, 4 May 2023 22:18:05,
David E. Ross <nobody@nowhere.invalid> writes
>On 5/4/2023 9:43 PM, J. P. Gilliver wrote:
>> In message <1341h75tghz67.dlg@v.nguard.lh> at Thu, 4 May 2023 22:35:46,
>> VanguardLH <V@nguard.LH> writes
>>> Can you connect to AV sites, like Norton, Avast or AVG, Avira,
>>> Bitdefender, etc?
>>
>> Do you mean can I connect to those companies' websites, or something
>> else?
>>
>
>I think it means you should try connecting by using an application that
>is NOT a browser or E-mail client. For example, my anti-virus is AVG

I've just got up, and (I left it on, as usual) looking at the news/mail
client log, I see it was connected from about 03:56:56 until 08:37:44
(yes, I sometimes sleep odd hours!) when it failed, and remained failed
(it is now 22:11). [Interestingly, that's about 4 hours 40 minutes to
failure - I _think_ it was that last time I looked. Any significance to
about 280 minutes?]

>AntiVirus Free; it connects directly to an AVG server several times a
>day to obtain the latest virus definitions. I also have the

Mine is AVG, and I've just looked: it says last updated 2 (now 3)
minutes ago. So _that's_ obviously still connecting!

>clock-synchronization application SocketWatch, set to query network time
>protocol (NTP) clocks hourly around the world to maintain the correct
>time for my PC clock. Finally, although updates to my Windows 7 ceased

I've just used the normal Windows time system (right-click on clock,
Adjust date/time, Internet Time tab, Change settings..., Update now
button [it's set to time.windows.com]): "The clock was successfully
synchronised with time.windows.com on ... at ..." - and again when I
tried three minutes later.

>some time ago, I still get virus definitions through Windows Update for
>Microsoft Security Essentials. All of these involve Internet
>connections directly and not through any browser or E-mail client.
>
Seems I have the same. Even within the browser, some things work -
twitter partially (I see tweets but not the pictures in them!), most
things don't. (Including YouTube and Google this time.)

netstat -nao | find /i "estab" /c

returns 3. yt-dlp fails.

ipconfig /flushdns

returns "Successfully flushed the DNS Resolver Cache.", but hasn't made
things work. Pinging news.eternal-september.org resolves to the same IP
as it does for the news client, and does four successful pings.
(Interestingly pinging news.plus.net - which actually pings
isp.news.ord.giganews.com - gives a different IP in the last quad, but
that comes out different every ping, so I presume it's a farm.)

All very curious!
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"Who came first? Adam or Eve?" "Adam of course; men always do."
Victoria Wood (via Peter Hesketh)

Posting - or rather typing, it won't leave the outbox - under "fail"
condition.

In message <6RC0pxGszIVkFwk3@255soft.uk> at Fri, 5 May 2023 05:58:20, J.
P. Gilliver <G6JPG@255soft.uk> writes
>In message <u31tnu$26nvm$1@dont-email.me> at Thu, 4 May 2023 23:38:38,
>Paul <nospam@needed.invalid> writes
[]
>> To check the total number of established connections, run this command:
>>
>> netstat -nao | find /i "estab" /c
>> "
>
>That says 22 at the moment, but the fault isn't present yet.

(From an Administrator command prompt) gave 3; when I tried again, gave
2. Three more tries gave 2.
>>
>>That's roughly the equivalent of using TCPView in a sense.
>>I'm not sure what we're supposed to see, by doing that.
>>
>>The netstat -nao doesn't seem right somehow.
>>
>>I tried netstat -s
>>to see Current Connections
>>
>> TCP Statistics for IPv4
>> Current Connections = 21
>
>I have 14 at the moment.

2 now. Four times in a row.
>>
>>But that doesn't tell me why I'm out of NonPaged Pool memory.
>>
>>If you're running Torrents in the background, any kind
>>of exhaustion (even in the router) is possible.
>>
>I'm pretty sure I'm not running Torrents, assuming that's something I
>would have started rather than being something the OS might do on its
>own.
>>*******
>>
>>There is PoolMon. Pools can be assigned a four character name
>>like "PIGZ", then it's up to you to figure out which software
>>you installed which is consuming NPP.
>>
>>https://learn.microsoft.com/en-us/windows-hardware/drivers/devtest/poolmon
>>
>>The thing is, on Windows XP, Pool memory is strictly limited.
>>Like maybe 256MB on a 3GB machine kind of limited.
>>
>>One change on Vista+ OSes, was Pool could use the entire
>>computer memory, or damn close to it. You won't constantly
>>running up against garbage collector behavior to get
>>pool to work with on Vista or later.

That looks a little above my comprehension level.
>>
>>The Control Panels has Troubleshooters, and there may
>>be one for networking. It contains the usual two commands
>>for resetting the network.

I've done Control Panel, Troubleshooting, Network and Internet. Five
choices: Internet Connections, Shared Folders, HomeGroup, Network
Adapter, and Incoming Connections. Trying the first one, it gets to
"Windows could not automatically detect this network's proxy settings"
"Detected" and a yellow triangle. Trying Network Adapter gets
"Troubleshooting couldn't identify the problem" (which seems OK, as I
don't think there's any hardware fault, certainly not with the wifi or
"router" [modem/router/wifi box]).
>>
>>However, this does not reset the Pool, so if something
>>is chowing down on Pool, you'll rapidly run into the same
>>issue again. I don't think you can escape, until you discover
>>what unrelated program is "leaking pool".

Sounds sensible.
>>
>>You might want to glance at Task Manager, and see if
>>the memory usage makes sense. If there seemed to be
>>no free memory left, maybe that's a hint too of an issue.
>
>Will do next time the fault arises. But I don't think it's something
>that is hogging general resources like memory or CPU, as there's no
>increase in fan speed or perceptible slowing - I can carry on doing
>whatever else I want, I don't become aware that _some_ of the
>connections to the outside world have failed until I actually try to
>use them.

Total memory in use is 2.26 - now 2.25 - GB, out of Total 3484 available
according to Task Manager's Performance tab (the machine has 4G fitted);
the Processes tab (with "Show processes from all users" ticked) shows
the top three processes as chrome.exe, all at 1xx,xxx K (there are other
chrome.exe processes); closing chrome drops the total used to 1.44 GB,
with the highest being svchost.exe at 75,492 K. Basically, the machine
doesn't seem stressed in that respect - or CPU; it's showing System Idle
Process hovering around 91 to 97, or CPU usage at 1 to 9 %.
>
>>Any time there's a leak (especially on an unbounded
>>consumer), you might see signs in Task Manager. They
>>won't be charged to Firefox, but to some system thing perhaps.
>
>The poor much-maligned svchost, perhaps!
>>
>> Paul
>John

Well, it is the top one, and there are 12 altogether (7 LOCAL SERVICE, 2
NETWORK SERVICE, 3 SYSTEM), but none of them seem to be hogging memory
or CPU.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"Who came first? Adam or Eve?" "Adam of course; men always do."
Victoria Wood (via Peter Hesketh)

In message <nxeuklusibbw$.dlg@v.nguard.lh> at Fri, 5 May 2023 14:09:05,
VanguardLH <V@nguard.LH> writes
>"J. P. Gilliver" <G6JPG@255soft.uk> wrote:
>
>> VanguardLH <V@nguard.LH> WROTE:
>>
>>> Can you connect to AV sites, like Norton, Avast or AVG, Avira,
>>> Bitdefender, etc?
>>
>> Do you mean can I connect to those companies' websites, or something
>> else?
>
>Yes, can you connect to those web sites. If you were infected, many
>malware will block you getting to AV sites. They don't want you going
>where you could eradicate them. Google and Youtube (also Google) don't
>have AV programs.

After the fault has cut in, I'd expect most websites to fail, though
I'll try a few if I remember. Now, before it has, I can access avg.com,
norton.com, avast.com, avira.com (at which point an AVG "We've safely
aborted ..." warning popped up!), and bidefender.com .
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Only dirty people need wash

In message <u33nic.kos.1@ID-201911.user.individual.net> at Fri, 5 May
2023 18:05:41, Frank Slootweg <this@ddress.is.invalid> writes
[]
> I haven't followed all the details in this thread, so I may be wrong,
>but I get the impression that if things don't work, the *only* thing
>which works is a browser, everything else fails.
>
> If that is correct, isn't it a problem with any port other than 80? If
>so, shouldn't you be looking at your firewall (and your antivirus
>software and your (built-in) router (did I forget anything?)) for port
>related problems?

No, it's the other way round - once the fault has cut in, most things
(i. e. web pages) _in_ the browser don't work (this last time, even
YouTube and Google [the search engine] didn't), outside some things -
ping, time server, AVG - still do work. Though some things outside -
yt-dlp, my news/email client - don't.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

If vegetarians eat vegetables,..beware of humanitarians!

In message <u32us1$qiq2$1@paganini.bofh.team> at Fri, 5 May 2023
09:03:23, Newyana2 <Newyana2@invalid.nospam> writes
>"J. P. Gilliver" <G6JPG@255soft.uk> wrote
>
>| >> Connection failed - Cannot connect
>| >> -- Winsock ERROR : No buffer space available
>| >
>
> In addition to Paul's info, there's a lot online about
>this error:
>
>https://stackoverflow.com/questions/16712354/windows-socket-error-code-10055

Reading that one, he says "the internet connection begin to fail and
finally I get this error (10055), which means that my app run out of
buffer space." Well, for me, I get no "begins to fail" - it's fine,
until suddenly is isn't. (Though it could be a threshold effect I
suppose.) Most of the followups seemed to be suggesting faults in the
OP's app - I've not written one. (My programming days are long gone,
apart from the odd trivial batch file.)
>
> I would wonder about the habit of leaving tabs open,
>especially if some may be reloading periodically. Websites
>these days are often very large software programs.
>
> A couple of possibilities (once you replace Chrome with
>a civilized browser :) would be to just stop leaving tabs

I'm aware of your redmist reflex when anyone mentions (a) Chrome (b)
multiple tabs, and I understand your reasons. Nevertheless, (a) I don't
think my version of Chrome has changed since before this problem started
happening [I get the "you need later than Windows 7 to get the latest
Chrome" message most times I start it, so presume I've always been on
the same version, unless the last version that worked with 7 was very
recent], and (b) I don't have many tabs open - 17 at the moment (make
that 16, I've just closed the above one), 3 + 2 + 4 of those to
different parts of the same sites, virtually none of them to sites with
any sort of refresh (I think Twitter does).

I don't think I've changed either my version of Chrome or the average
number of tabs I have open since before this problem started to appear;
certainly I'm pretty sure that opening more tabs doesn't trigger it
(see below under YouTube).

>open and to stop reloads. In FF that's Accessibility.blockautorefresh*
>It may be available in Chrome. I don't know the secret nooks
>and crannies of Chrome, or how much it lets you control.
>You can also disable prefetch. The waste in browsers these
>days is astonishing... and senseless. They cut corners on

I certainly agree with you there.

>images to make them smaller and delay loading, then they send
>you to get 15 MB of script and 12 other webpages to load
>in the background.

Well, I don't have the background bit (AFAIK), but I agree
wholeheartedly about the script: they take MB to do what used to be done
in a few hundred B. And, quite often, it doesn't work properly anyway -
in particular, header, footer, and side parts, which obscure too much of
the page, and text that forces you to scroll (up/down and/or side/side),
when raw HTML _intrinsically_ autowraps unless you code to deliberately
stop it doing so.
>
>* News sites, for example, often reload every few minutes.
>I found that maddening, to have a page reload while I'm
>reading it. So I blocked reloads. If you have 50 tabs open

Oh yes, YouTube does that - I get a reformat when I'm typing a comment,
when the clip ends and an advert loads that has a different aspect
ratio. Having said that, I sometimes have lots of YouTube tabs open in
addition to my normal ones - if I'm viewing a YT clip and see another
one that looks interesting, I open it in a new tab (I don't use their
"playlist" thing), then view the clips, closing each one as I finish
with it (or before if it's not interesting), and having lots of those
open _doesn't_ seem to trigger the problem: in fact I can carry on
viewing and closing the ones I've opened after the fault has cut in, not
being aware it has done so unless I try to yt-dlp one of them. I think
I've maybe had getting on for 20 of those open.

>you could very well have almost constant reloads happening.
>
Agreed. Don't think I've got anywhere close to 50 - normally about what
I have now, apart from sometimes the YouTube ones which get closed the
same day.
>
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

If vegetarians eat vegetables,..beware of humanitarians!

On 5/5/2023 5:59 PM, J. P. Gilliver wrote:

>
> Well, it is the top one, and there are 12 altogether (7 LOCAL SERVICE, 2 NETWORK SERVICE, 3 SYSTEM), but none of them seem to be hogging memory or CPU.

OK, Chrome 109 has B-Y-O-B buffer Webtransport. This
was intended to support zero copy operation with
networking. B-Y-O-B stands for bring your own buffer,
meaning a Ring 3 process offers a buffer to a Ring 0 driver.
Otherwise, how could you offer a buffer, if application
memory comes from Ring 3 ?

Normal network stacks might copy blocks of memory (buffers)
from one layer in the stack to another. A zero copy strategy,
means an application offers a pointer for a buffer, and the
network driver puts that buffer in the ring buffer. Something
along those lines.

There have been other ways to do this. Programmable DMA in the
past (like breaking a packet into five separate pieces and
storing them individually), was a way to avoid block copies.

However, the network driver is unlikely to be aware of what
you're doing.

With your symptoms, my concern would be there is some sort
of memory fragmentation issue caused by screwing around
with the network stack.

It does not appear this feature is in Firefox. Chrome 109 is
the last version of Chrome for Windows 7 SP1.

# Sample non-Googlean documentation
https://developer.mozilla.org/en-US/docs/Web/API/WebTransport

Initially, I was looking for a Chrome flag to turn this
off, but searches are useless. Nothing remotely looking like
a control is coming back in a search.

1) Is your lack of functional connections only in Chrome ?
2) Is your version of Chrome 109 or something else ?
2) At the point that there are connection failures, does
Firefox always work or not ?

I was trying to find a Chrome download straight from Google, but
all I can find so far are references to flim-flam sites
offering copies of Chrome. For which they used to earn $1 each.
I generally try to get installs from their supplier if
I can manage it. With Firefox, I use use their release server.

This is highly unlikely to be in usage today, but what do I know.
I'm not exactly interested in web evolution, because it has
nothing to do with end users, and everything to do with
greasing business skids.

Paul

"J. P. Gilliver" <G6JPG@255soft.uk> wrote

|
>https://stackoverflow.com/questions/16712354/windows-socket-error-code-10055
| | Reading that one, he says "the internet connection begin to fail and
| finally I get this error (10055), which means that my app run out of
| buffer space." Well, for me, I get no "begins to fail" - it's fine,
| until suddenly is isn't. (Though it could be a threshold effect I
| suppose.) Most of the followups seemed to be suggesting faults in the
| OP's app - I've not written one.

Yes, but it's helpful to know the actual error number.
Errors can have different causes, but the exhausted buffer
aspect implies that somewhere a limit is being reached.
Unfortunately I don't know enough about it, but winsock
is the library that handles actual internet communication.
They're talking about an apparent memory allotment per-
program. But there could also be a winsock instance limit
or memory use limit for the system. Just a guess. That's
why I suggested the various efforts that might reduce
the load.

| I'm aware of your redmist reflex

"Red mist". I had to look that up. I guess it must refer to
blood in the eyeball during extreme anger. Like the idea
of "seeing red". I should remember that you have a low
tolerance for ribbing, wisecracks, and general, artful
negativity. :)

I've just been watching the coronation, here in the colonies.
Amazing piece of work. I wish I could have seen an up-close,
extended look at the gold carriage decorations. But at least
the announcers were quite competent about explaining details
that I would have otherwise missed. I find it a relief to see the
value of ritual recognized, even if it might seem out of date.

Just had another fail - about 3h30 this time. So 210-220 minutes seems
to be the uptime. Any thoughts?
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

There's too much attention paid to how TV can be bad for you, but I think it's
good for us more often than it's bad - Professor Barrie Gunter of Sheffield
University (quoted in RT, 15-21 March 2003).

In message <u34nr9$2nsvu$1@dont-email.me> at Sat, 6 May 2023 01:16:23,
Paul <nospam@needed.invalid> writes
>On 5/5/2023 5:59 PM, J. P. Gilliver wrote:
>
>> Well, it is the top one, and there are 12 altogether (7 LOCAL
>>SERVICE, 2 NETWORK SERVICE, 3 SYSTEM), but none of them seem to be
>>hogging memory or CPU.
>
>OK, Chrome 109 has B-Y-O-B buffer Webtransport. This
[]
>Normal network stacks might copy blocks of memory (buffers)
[]
>There have been other ways to do this. Programmable DMA in the
[]
>However, the network driver is unlikely to be aware of what
>you're doing.

All way above my head, I'm afraid!
[]
>1) Is your lack of functional connections only in Chrome ?

No - my email/news client loses connection to the mail and news servers;
it was actually from that client's log window that I discovered the "out
of buffer space" hint - something like "failed to connect to <> server,
not enough buffer space". Also yt-dlp ceases to work (which is called
from a command prompt). AVG seemed to be continuing to get updates, and
ping - from a command prompt - worked (it resolved an alphabetic address
to a numeric IP, then did the usual four pings which succeeded).

>2) Is your version of Chrome 109 or something else ?

"Version 109.0.5414.120 (Official Build) (32-bit)".

>2) At the point that there are connection failures, does
> Firefox always work or not ?

Hadn't thought to try it; will if I remember next fail. (Assuming
consistency, that'll be in about 3 hours.)
>
>I was trying to find a Chrome download straight from Google, but
>all I can find so far are references to flim-flam sites
>offering copies of Chrome. For which they used to earn $1 each.
>I generally try to get installs from their supplier if

Oh, definitely. (Though oldversion.com is reasonable.)

>I can manage it. With Firefox, I use use their release server.

Used to even be an ftp rather than http server; haven't looked for a
while.
>
>This is highly unlikely to be in usage today, but what do I know.
>I'm not exactly interested in web evolution, because it has
>nothing to do with end users, and everything to do with
>greasing business skids.

Indeed. Although I do have the two browsers - Firefox and Chrome (and I
think IE is still there) - on this machine, I have little interest in
developments. (If anything, I'd probably give a brief look at things
that _block_ a lot of such, but am losing the will to fight.)
>
> Paul
John
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

There's too much attention paid to how TV can be bad for you, but I think it's
good for us more often than it's bad - Professor Barrie Gunter of Sheffield
University (quoted in RT, 15-21 March 2003).

In message <u35j4a$17v2a$1@paganini.bofh.team> at Sat, 6 May 2023
09:01:23, Newyana2 <Newyana2@invalid.nospam> writes
>"J. P. Gilliver" <G6JPG@255soft.uk> wrote
>
>|
> >https://stackoverflow.com/questions/16712354/windows-socket-error-code-10055
>|
>| Reading that one, he says "the internet connection begin to fail and
>| finally I get this error (10055), which means that my app run out of
>| buffer space." Well, for me, I get no "begins to fail" - it's fine,
>| until suddenly is isn't. (Though it could be a threshold effect I
>| suppose.) Most of the followups seemed to be suggesting faults in the
>| OP's app - I've not written one.
>
> Yes, but it's helpful to know the actual error number.

Assuming it's the same one, yes. The symptom description is near enough
(to what my news/email client log window said) that it probably is, but
the log window did not actually give an error number.

>Errors can have different causes, but the exhausted buffer
>aspect implies that somewhere a limit is being reached.

Indeed. Just based on two tries, it seems it might be time-connected
(210-220 minutes) rather than number of actions, but that might not be
the case - we'll see this time.

>Unfortunately I don't know enough about it, but winsock
>is the library that handles actual internet communication.
>They're talking about an apparent memory allotment per-
>program. But there could also be a winsock instance limit
>or memory use limit for the system. Just a guess. That's

It does seem somewhat likely that something is running out, yes. (Not
memory or CPU time.)

>why I suggested the various efforts that might reduce
>the load.

I tried some of them - see earlier posts - with, sadly, no success.
>
>| I'm aware of your redmist reflex
>
> "Red mist". I had to look that up. I guess it must refer to
>blood in the eyeball during extreme anger. Like the idea
>of "seeing red". I should remember that you have a low
>tolerance for ribbing, wisecracks, and general, artful
>negativity. :)

I may have made up the word as a single word, but yes, it does refer to
a loss of rational thought when triggered by certain things. I
remembered from of old that mention of multiple tabs, or chrome, used to
sometimes trigger a "mayayana rant" (I assume you're the Mayayana of
old), which would go off at length, forgetting what we're actually
discussing. (You didn't do that really this time.)
>
> I've just been watching the coronation, here in the colonies.
>Amazing piece of work. I wish I could have seen an up-close,
>extended look at the gold carriage decorations. But at least
>the announcers were quite competent about explaining details
>that I would have otherwise missed. I find it a relief to see the
>value of ritual recognized, even if it might seem out of date.
>
Yes, I've watched it - it's been more or less all day here on BBC1.
Though not an ardent royalist, I'm more pro than anti - it's rather fun
having such an archaic institution, and despite their not inconsiderable
cost, I would say I get more than good value out of them - plus they do
quite a lot for our tourism budget. Yes, it all - the service and the
various parades - went off without a hitch: we _do_ do that sort of
thing very well. (I did think Charles looks old though - which is fair
enough, he is!)

(Like most Brits, I don't give they royals a moment's thought most of
the time. Germany has something similar in their president [as opposed
to chancellor]: a mostly ceremonial head. I can't think of anything
similar the USA has.)
>
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

There's too much attention paid to how TV can be bad for you, but I think it's
good for us more often than it's bad - Professor Barrie Gunter of Sheffield
University (quoted in RT, 15-21 March 2003).

On 5/6/2023 2:23 PM, J. P. Gilliver wrote:
> In message <u34nr9$2nsvu$1@dont-email.me> at Sat, 6 May 2023 01:16:23, Paul <nospam@needed.invalid> writes

>> 2) Is your version of Chrome 109 or something else ?
>
> "Version 109.0.5414.120 (Official Build) (32-bit)".

Just FYI. A Win7 version (more or less, end of support).

Using your version for bootstrap, I get a search engine and download
locations which look reasonable. I usually tag the downloads a bit,
for later consumption.

https://chromium.cypress.io/ # These items marked as "Stable"

https://chromium.cypress.io/win64/stable/109.0.5414.120
https://www.googleapis.com/download/storage/v1/b/chromium-browser-snapshots/o/Win_x64%2F1070094%2Fchrome-win.zip?generation=1668133267247822&alt=media

Name: Win_x64_1070094_chrome-win__109.0.5414.120-x64.zip
Size: 201,716,092 bytes (192 MiB)
SHA256: 340292ED751222AFD2293FD5A406E564EE391144EB1E269FB3AAA6D91AB2ACE1

https://chromium.cypress.io/win/stable/109.0.5414.120
https://www.googleapis.com/download/storage/v1/b/chromium-browser-snapshots/o/Win%2F1070113%2Fchrome-win.zip?generation=1668135846921518&alt=media

Name: Win_1070113_chrome-win____109.0.5414.120-x86.zip
Size: 188,904,887 bytes (180 MiB)
SHA256: 60552F33B0CDF872ACC76010517ACEB3C2178CF37B4D38017770C1B8DDEFC691

Paul

In message <u36haq$30uve$1@dont-email.me> at Sat, 6 May 2023 17:37:29,
Paul <nospam@needed.invalid> writes
>On 5/6/2023 2:23 PM, J. P. Gilliver wrote:
>> In message <u34nr9$2nsvu$1@dont-email.me> at Sat, 6 May 2023
>>01:16:23, Paul <nospam@needed.invalid> writes
>
>>> 2) Is your version of Chrome 109 or something else ?
>> "Version 109.0.5414.120 (Official Build) (32-bit)".
>
>Just FYI. A Win7 version (more or less, end of support).

That's what I thought.
[]
The only installer file I can find on my machine that I would have
downloaded (i. e. on D:) is ChromeStandaloneSetup [32 bit] dated (in my
machine, so probably when I downloaded it) 2023-1-28, size 91,890,728 (I
would have added the " [32 bit]" to the filename); the chrome.exe is
dated 2023-1-24 (and 2,465,048, in C:\Program
Files\Google\Chrome\Application\). There's also chrome_installer.exe in
C:\Program
Files\Google\Update\Download\{8A69D345-D564-463C-AFF1-A69D9E530F96}\109.0
..5414.120\, dated 2023-1-28, 89,268,264, which I wouldn't have fetched
but chrome itself (an earlier version?) might have.

I don't _think_ it's Chrome that's the cause - but it could be I
suppose, if it's got some sort of leak.

The time to fail this time was about (log file info) 18:46:29 to
23:16:34, or about 4½ hours.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

If it ain't broke, fix it 'til it is.

On 5/6/2023 6:53 PM, J. P. Gilliver wrote:
> I don't _think_ it's Chrome that's the cause - but it could be I suppose, if it's got some sort of leak.
>
> The time to fail this time was about (log file info) 18:46:29 to 23:16:34, or about 4½ hours.

Would there be anything running in the background ?

An email program checking for "new mail" every once in a while ?

If I use

about:serviceworkers

in Firefox, it says

Service Workers are not enabled

and that's an example of a background service.

And we have the piece of evidence in your log, as
the only solid lead.

"Socket Error 10055 - No Buffer Space Available

Error 10055 means that Windows has run out of TCP/IP socket buffers
because too many connections are open at once.

Things to check if this happens are:

Is there any other TCP/IP software running on the same PC as VPOP3 - eg a proxy server
Use a virus scanner to check for malware, as some malware is badly written and will use a lot of connections
Use a program such as TCPView or NETSTAT (with the -b parameter) to see which programs have connections open
"

But that's not the only mechanism. Pool memory exhaustion
is another mechanism.

Paul

In message <u36ruj$32dck$1@dont-email.me> at Sat, 6 May 2023 20:38:42,
Paul <nospam@needed.invalid> writes
>On 5/6/2023 6:53 PM, J. P. Gilliver wrote:
>> I don't _think_ it's Chrome that's the cause - but it could be I
>>suppose, if it's got some sort of leak.
>> The time to fail this time was about (log file info) 18:46:29 to
>>23:16:34, or about 4½ hours.
>
>Would there be anything running in the background ?

That does seem the most likely cause of the problem.
>
>An email program checking for "new mail" every once in a while ?

The news/mail client does indeed do that, but the problem had arisen
before I installed the client.
>
>If I use
>
> about:serviceworkers
>
>in Firefox, it says
>
> Service Workers are not enabled

(Sounds like a cue for industrial action - sorry, we've got a lot of
strikes here at the moment!)

Let me see: Firing up Firefox ... interesting! I have three of them,
none of which I see as needing to be on continuously (for interest, they
are: https://www.ancestry.co.uk/dna/origins/service-worker.js,
https://speedtest.btwholesale.com/service-worker.js, and
https://www.vistax64.com/service_worker.js - I don't even know what the
last one _is_). Each has an "Unregister" button; I've just clicked
those, and it now says

No Service Workers registered.

So we'll see. Sounds like a useful tip, anyway.
>
>and that's an example of a background service.
>
>And we have the piece of evidence in your log, as
>the only solid lead.
>
> "Socket Error 10055 - No Buffer Space Available
>
> Error 10055 means that Windows has run out of TCP/IP socket buffers
> because too many connections are open at once.
>
> Things to check if this happens are:
>
> Is there any other TCP/IP software running on the same PC as VPOP3
>- eg a proxy server

I'm pretty sure I'm not using, providing, or anything else, any proxy.

> Use a virus scanner to check for malware, as some malware is badly
>written and will use a lot of connections

When I go to bed I'll set AVG going on a scan, if I remember.

> Use a program such as TCPView or NETSTAT (with the -b parameter) to
>see which programs have connections open
> "
I suspect those are beyond me.
>
>But that's not the only mechanism. Pool memory exhaustion
>is another mechanism.
>
> Paul
John
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

I'd rather trust the guys in the lab coats who aren't demanding that I get up
early on Sundays to apologize for being human.
-- Captain Splendid (quoted by "The Real Bev" in mozilla.general, 2014-11-16)

In message <J564j7DNSxVkFwR6@255soft.uk> at Sun, 7 May 2023 04:01:33, J.
P. Gilliver <G6JPG@255soft.uk> writes
>In message <u36ruj$32dck$1@dont-email.me> at Sat, 6 May 2023 20:38:42,
>Paul <nospam@needed.invalid> writes
>>On 5/6/2023 6:53 PM, J. P. Gilliver wrote:
>>> I don't _think_ it's Chrome that's the cause - but it could be I
>>>suppose, if it's got some sort of leak.
>>> The time to fail this time was about (log file info) 18:46:29 to
>>>23:16:34, or about 4½ hours.

Latest one 19:26:09 to 23:56:17 - almost exactly 4½ hours again.
>>
>>Would there be anything running in the background ?
>
>That does seem the most likely cause of the problem.
>>
>>An email program checking for "new mail" every once in a while ?
>
>The news/mail client does indeed do that, but the problem had arisen
>before I installed the client.
>>
>>If I use
>>
>> about:serviceworkers
>>
>>in Firefox, it says
>>
>> Service Workers are not enabled
>
>(Sounds like a cue for industrial action - sorry, we've got a lot of
>strikes here at the moment!)
>
>Let me see: Firing up Firefox ... interesting! I have three of them,
>none of which I see as needing to be on continuously (for interest, they
>are: https://www.ancestry.co.uk/dna/origins/service-worker.js,
>https://speedtest.btwholesale.com/service-worker.js, and
>https://www.vistax64.com/service_worker.js - I don't even know what the
>last one _is_). Each has an "Unregister" button; I've just clicked
>those, and it now says
>
> No Service Workers registered.
>
>So we'll see. Sounds like a useful tip, anyway.

Tried it again after latest fail - "No Service Workers registered."
still.
[]
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

>
> Latest one 19:26:09 to 23:56:17 - almost exactly 4� hours again.
>>>

Have you considered that the router/modem might be malfunctioning?
Sometimes old routers display tantrums and no matter how many times you
bang your head on a concrete wall, it won't budge.

If it is an ISP provided router then phone them for a replacement and if
they refuse tell them you want to move to some other corporation who
will provide a new free router. They will just give you one free if you
threaten them.

In message <u39j5o$1ufhs$1@paganini.bofh.team> at Mon, 8 May 2023
02:24:14, Ramos <invalid@invalid.net> writes
>
>>
>> Latest one 19:26:09 to 23:56:17 - almost exactly 4½ hours again.
>>>>
>
>Have you considered that the router/modem might be malfunctioning?
>Sometimes old routers display tantrums and no matter how many times you
>bang your head on a concrete wall, it won't budge.

Router is a recent model, received January or later; the 4½ hours is
from when I restart the PC (the router is on continuously, showing
blue).
>
>If it is an ISP provided router then phone them for a replacement and if
>they refuse tell them you want to move to some other corporation who
>will provide a new free router. They will just give you one free if you
>threaten them.

I am pretty convinced it is a PC software fault.
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

Twitter makes you despair of people you don't know, and Facebook makes you
despise people you do know. - Lucy Porter on @Room 101@, 2016

On 5/8/2023 5:11 AM, J. P. Gilliver wrote:
> In message <u39j5o$1ufhs$1@paganini.bofh.team> at Mon, 8 May 2023 02:24:14, Ramos <invalid@invalid.net> writes
>>
>>>
>>> Latest one 19:26:09 to 23:56:17 - almost exactly 4½ hours again.
>>>>>
>>
>> Have you considered that the router/modem might be malfunctioning?
>> Sometimes old routers display tantrums and no matter how many times you
>> bang your head on a concrete wall, it won't budge.
>
> Router is a recent model, received January or later; the 4½ hours is from when I restart the PC (the router is on continuously, showing blue).
>>
>> If it is an ISP provided router then phone them for a replacement and if
>> they refuse tell them you want to move to some other corporation who
>> will provide a new free router. They will just give you one free if you
>> threaten them.
>
> I am pretty convinced it is a PC software fault.

You might be able to tell, by testing of a second
computer while the first one is showing failure signs.

If the router is borked, then both computers would suffer
service issues at the same time.

The "10055: No buffer space" recorded on the affected laptop,
is proof it's an OS issue.

Paul

J. P. Gilliver <G6JPG@255soft.uk> wrote:
[...]

> I am pretty convinced it is a PC software fault.

Another long shot:

Have you updated the driver for your network adapter (WiFi? LAN?)
recently, either intentionally or 'something' did it behind your back?

Recently Windows 11 pulled the 'behind your back' stunt and that gave
*very* weird disconnect errors. After rolling back the WiFi driver,
everything was/is fine.

A driver problem does not really match with the "10055: No buffer
space" error, but you never know.

On Wed, 3 May 2023 21:24:08 +0100, "J. P. Gilliver" <G6JPG@255soft.uk>
wrote:

>W7-32; I don't know if that's relevant.
>
>This machine loses _some_ connection after some hours; a restart cures
>it. It only started happening in the last month or so, not from when I
>got it, so it's something I've done - unfortunately I don't know what.
>
>After whatever happens has happened: Many - probably most - websites
>don't work (e. g. BBC Weather, Twitter, Ancestry). Ping works (so I've
>[a] got _some_ connection [b] got DNS). yt-dlp fails (with a page full
>of error messages). BUT: YouTube and Google (the search engine)
>_continue to work_; I can view clips I haven't viewed before, and add
>comments, on YouTube, and I can do a Google search I haven't done
>before, and get pages of results (though mostly, if not entirely,
>clicking on the links in them doesn't work).
>
>Any idea what's going on, and/or where I should look? It seems a very
>odd fault!
>
>Yes, it hasn't escaped my notice that YouTube and Google (and the Chrome
>browser I'm using - latest for W7 I think) are all from the same
>company.

(Coming to this late, but...)

I had a similar issue; while surfing, I will suddenly "lose
connection" to the network. Any new tabs I open will - after a while -
time-out, and attempting to check my email or basically do anything on
the Internet (Steam, SFTP) will silently fail. Rebooting would solve
the problem, but that always felt like overkill.

Fortunately, there was a simpler solution: killing the Firefox
executable. And by killing it, I mean more than just closing all the
open Firefox windows; I had to go into Task Manager and make sure each
and every firefox.exe instance was terminated and gone. Once that
happened, connectivity was restored as if nothing happened. Firefox
worked, email worked, everything worked. But until that point...

I don't really know if it was a problem with Firefox, some add-on, an
underlying OS issue, or the router. My uneducated guess was that
Firefox was oversaturating the router (or OS?) with requests and the
thing just hung until I got rid of it. But even if it only happened
when I was /using/ Firefox, the effect wasn't just /limited/ to
Firefox; if I tried Chromium, or my email client, I'd get network
time-outs until I killed firefox.exe

I'm not sure if this will be helpful, but there you go...

In message <u3ahhp$3rkrc$1@dont-email.me> at Mon, 8 May 2023 06:05:45,
Paul <nospam@needed.invalid> writes
[]
>You might be able to tell, by testing of a second
>computer while the first one is showing failure signs.

That would indeed be a good way to eliminate the router. Unfortunately I
don't have another computer ...
>
>If the router is borked, then both computers would suffer
>service issues at the same time.
>
>The "10055: No buffer space" recorded on the affected laptop,
>is proof it's an OS issue.
>
> Paul
>
.... and I suspected it was in this computer, anyway. (I've not seen the
"10055", only the "No buffer space" - the number came from someone in
this thread who knew of an article connecting the number to the
description.)
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"The great tragedy of science, the slaying of a beautiful theory by an ugly
fact. - Thomas Henry Huxley

In message <u3b50n.bls.1@ID-201911.user.individual.net> at Mon, 8 May
2023 13:39:57, Frank Slootweg <this@ddress.is.invalid> writes
>J. P. Gilliver <G6JPG@255soft.uk> wrote:
>[...]
>
>> I am pretty convinced it is a PC software fault.
>
> Another long shot:
>
> Have you updated the driver for your network adapter (WiFi? LAN?)
>recently, either intentionally or 'something' did it behind your back?

Certainly not consciously, and I wouldn't have _thought_ Windows 7 would
have done it (though _something_ else might have I suppose). I maybe
_could_ eliminate the wifi link by connecting the laptop to the "router"
by cable rather than wifi. I might still try that next time ...
>
> Recently Windows 11 pulled the 'behind your back' stunt and that gave
>*very* weird disconnect errors. After rolling back the WiFi driver,
>everything was/is fine.
>
> A driver problem does not really match with the "10055: No buffer
>space" error, but you never know.

.... except that that wouldn't explain why _some_ things (ping, AVG,
_some_ web pages) still work after it "goes wrong".

Having said that: I've just noticed it's not been "up" for over 9¼
hours, so the 4½ hours the last few fails have been _isn't_ always the
case. Unless it's "got better" somehow!
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"The great tragedy of science, the slaying of a beautiful theory by an ugly
fact. - Thomas Henry Huxley

In message <i20i5itj6vjunpb1j4kvllh6hm2stc5if0@4ax.com> at Mon, 8 May
2023 10:08:11, Spalls Hurgenson <spallshurgenson@gmail.com> writes
[]
>(Coming to this late, but...)

The more the merrier!
>
>I had a similar issue; while surfing, I will suddenly "lose
>connection" to the network. Any new tabs I open will - after a while -
>time-out, and attempting to check my email or basically do anything on

The weird thing is that when it has "failed", _some_ things still have
connectivity - ping, AVG, _some_ web tabs. Like you, email/news
_doesn't_ work.

>the Internet (Steam, SFTP) will silently fail. Rebooting would solve
>the problem, but that always felt like overkill.

Cures it here too; had I been on my previous machine, I'd hate doing it
too, as that took ages to do a reboot, but this machine (no SSD) goes
from shutdown to back up in about 3 minutes (2 down, 1 up), so - since
the fault had been only occurring after about 4½ hours - it wasn't that
great an inconvenience: I'd go to the loo, get a coffee, or similar, and
it'd usually be back before I got back to the laptop.

As I said in my last post, it's now been "up" for getting on for 9½
hours, so I'm _hoping_ it's "got better", though I can't think what I've
done that's fixed it!
>
>Fortunately, there was a simpler solution: killing the Firefox
>executable. And by killing it, I mean more than just closing all the
>open Firefox windows; I had to go into Task Manager and make sure each
>and every firefox.exe instance was terminated and gone. Once that
>happened, connectivity was restored as if nothing happened. Firefox
>worked, email worked, everything worked. But until that point...
>
>I don't really know if it was a problem with Firefox, some add-on, an
>underlying OS issue, or the router. My uneducated guess was that
>Firefox was oversaturating the router (or OS?) with requests and the
>thing just hung until I got rid of it. But even if it only happened
>when I was /using/ Firefox, the effect wasn't just /limited/ to
>Firefox; if I tried Chromium, or my email client, I'd get network
>time-outs until I killed firefox.exe
>
>I'm not sure if this will be helpful, but there you go...
>
I fear not, as I don't usually have Firefox running. One could guess it
might be a similar fault with Chrome; I think I've tried stopping and
restarting that without success, though I haven't checked Task Manager
for lots of Chrome processes. I'll maybe look if the fault recurs.
>
>
>
Using Windows 7 32-bit, the last Chrome that works with it, and - when I
do use it (just firing it up now to see) Firefox "102.10.0esr (32 bit)"
(which also, I notice, says "Firefox is up to date" and "You are
currently on the esr update channel.", so obviously Firefox - up to the
last esr, anyway - is still supporting 7-32).
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"The great tragedy of science, the slaying of a beautiful theory by an ugly
fact. - Thomas Henry Huxley

In message <PKSJvJFg4TWkFwr+@255soft.uk> at Mon, 8 May 2023 19:23:28, J.
P. Gilliver <G6JPG@255soft.uk> writes
[]
>Having said that: I've just noticed it's not been "up" for over 9¼
>hours, so the 4½ hours the last few fails have been _isn't_ always the
>case. Unless it's "got better" somehow!

"now", not "not". Still "up" ...
--
J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

"The great tragedy of science, the slaying of a beautiful theory by an ugly
fact. - Thomas Henry Huxley

J. P. Gilliver <G6JPG@255soft.uk> wrote:
> In message <u3ahhp$3rkrc$1@dont-email.me> at Mon, 8 May 2023 06:05:45,
> Paul <nospam@needed.invalid> writes
> []
> >You might be able to tell, by testing of a second
> >computer while the first one is showing failure signs.
>
> That would indeed be a good way to eliminate the router. Unfortunately I
> don't have another computer ...

This remark triggered me to realize that (I think) you're currently
being forged in the comp.mobile.android newsgroup.

N.B. I think another poster (w....t) is being forged as well.

Sad state of affairs.

[...]