On Sat, 27 Sep 2008 22:47:41 +1000, Rob wrote:
> I have recently upgraded to version 2008.1 and have noticed an apparent
> difference between this version and 2007.1 which was the previous version I
> used.
>
> With the earlier version, if the automatic mandriva update applet showed
> updates were available, I could then use the 'urpmi --auto-update' command
> and it would find and install the updates. With 2008.1 if the applet
> indicates there are updates and I then run the manual update *no* updates
> are found. However if I then use the applet to perform the installation
> there are indeed updates available.


If you are going to run from the command line, use something like
urpmi --wget --update --auto-update --auto

That will verify your rpm lists match the mirror's list.
If different, download the new lists used by auto-update.
Just running with --auto-update only checks your rpm list against your
install.

I suggest you use the --wget option. Some sites only allow one
connection at a time from the same ip. If curl runs instead of wget,
you could miss some files.

Rob wrote:

> I have recently upgraded to version 2008.1 and have noticed an apparent
> difference between this version and 2007.1 which was the previous version
> I used.
>
> With the earlier version, if the automatic mandriva update applet showed
> updates were available, I could then use the 'urpmi --auto-update' command
> and it would find and install the updates. With 2008.1 if the applet
> indicates there are updates and I then run the manual update *no* updates
> are found. However if I then use the applet to perform the installation
> there are indeed updates available.
>
> Can someone please explain what is going on.
>
> I'm not sure if the following are relevant, but I have an entry in the
> urpmi.skip file (jpilot) AND I used the new easyurpmi system to set up the
> various repositories for 2008.1
>
> TIA
>
> Rob.


" urpmi --auto-update " == " urpmi.update -a && urpmi --auto-select "

I'm quite sure you used " urpmi --auto-select " instead of "
urpmi --auto-update " or, you haven't added an update media with
drakrpm-update.

This is an excellent tip - unless you try this from the command line you
probably wont get what it is this thing does for you.
Thanks Bit-Twister!
I learn something new every day :-)
Eric

[snip]
>
> If you are going to run from the command line, use something like
> urpmi --wget --update --auto-update --auto
>
> That will verify your rpm lists match the mirror's list.
> If different, download the new lists used by auto-update.
> Just running with --auto-update only checks your rpm list against your
> install.
[snip]

Bit Twister wrote:
> If you are going to run from the command line, use something like
> urpmi --wget --update --auto-update --auto
>
> That will verify your rpm lists match the mirror's list.
> If different, download the new lists used by auto-update.
> Just running with --auto-update only checks your rpm list against your
> install.
>
> I suggest you use the --wget option. Some sites only allow one
> connection at a time from the same ip. If curl runs instead of wget,
> you could miss some files.

The above does not seem to work properly when the repository
is accessed by rsync. For example, the following 3 lines appear
in a single burst, requiring a fraction of a second.

rsync://carroll.cac.psu.edu/mandrakelinux/official/2008.1/x86_64/media/non-free/backports/media_info/synthesis.hdlist.cz
....retrieving failed: rsync failed: exited with 10

problem reading synthesis file of medium "Non-free Backports"

On the other hand, the following works fine for me.
urpmi --wget --auto-update --auto-select --auto

Mandriva 2008.1
Linux localhost 2.6.25.11-tmb-desktop-2mdv #1 SMP Mon Jul 21 08:03:02
EDT 2008 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 5000+ GNU/Linux

Cheers!

jim b.

--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

On Sat, 27 Sep 2008 23:19:31 -0400, Jim Beard wrote:

> rsync://carroll.cac.psu.edu/mandrakelinux/official/2008.1/x86_64/media/non-free/backports/media_info/synthesis.hdlist.cz
> ...retrieving failed: rsync failed: exited with 10

That's because carroll.cac.psu.edu is currently refusing connections.

I'm currently getting ...
ftp://carroll.cac.psu.edu/pub/linux/...esis.hdlist.cz
....retrieving failed: wget failed: exited with 1

and ...
# ftp carroll.cac.psu.edu
ftp: connect: Connection refused

Regards, Dave Hodgins

--
Change nomail.afraid.org to ody.ca to reply by email.
(nomail.afraid.org has been set up specifically for
use in usenet. Feel free to use it yourself.)

David W. Hodgins wrote:
> On Sat, 27 Sep 2008 23:19:31 -0400, Jim Beard wrote:
>
>> rsync://carroll.cac.psu.edu/mandrakelinux/official/2008.1/x86_64/media/non-free/backports/media_info/synthesis.hdlist.cz
>> ...retrieving failed: rsync failed: exited with 10
>
> That's because carroll.cac.psu.edu is currently refusing connections.
>
> I'm currently getting ...
> ftp://carroll.cac.psu.edu/pub/linux/...esis.hdlist.cz
> ...retrieving failed: wget failed: exited with 1
>
> and ...
> # ftp carroll.cac.psu.edu
> ftp: connect: Connection refused

I ran the first line I listed, got the error message.

Ran the second line I listed, all worked and all my packages were up
to date.

Ran the first line again, so I could cut and paste the three lines
I wished to cite, and got the same error messages I had earlier.

Cheers!

jim b.


--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

On Sat, 27 Sep 2008 23:19:31 -0400, Jim Beard wrote:

> Bit Twister wrote:
>>
>> I suggest you use the --wget option. Some sites only allow one
>> connection at a time from the same ip. If curl runs instead of wget,
>> you could miss some files.
>
> The above does not seem to work properly when the repository
> is accessed by rsync. For example, the following 3 lines appear
> in a single burst, requiring a fraction of a second.
>
> rsync://carroll.cac.psu.edu/mandrakelinux/official/2008.1/x86_64/media/non-free/backports/media_info/synthesis.hdlist.cz
> ...retrieving failed: rsync failed: exited with 10
>
> problem reading synthesis file of medium "Non-free Backports"
>
> On the other hand, the following works fine for me.
> urpmi --wget --auto-update --auto-select --auto

Heheh, carroll is my favorite site for the single connect/ip failure.
I have a script which reads /etc/urpmi/urpmi.cfg and pull those directory's
package names, dates, size into an .idx file. Much faster greping *.idx
than using a urpm tool to find a package name. :-)

I had several retrieving failed messages, so I put a while loop waiting for
the last wget connection to drop before fetching the next directory.

Playing with cooker loads, from the command line, gives me the idea
that you might have drop outs depending on the size of the packages
being installed/removed.

I can guess your backports rsync was caused by the previous directory
having something or not downloaded. Unsure which.

I have taken to running the
urpmi --wget --auto-update --auto-select --auto
until I get all packages are up to date.

PS: carroll has been down since about 4pm Saturday. Guessing maintenance.

I have recently upgraded to version 2008.1 and have noticed a lost function
in the upgrade. I'm running Thunderbird as my mail client, and Opera 9.52
as my browser. In my browser, I can click an email address, and thunderbird
will open, but I can't click a link in Thunderbird and get Opera to open. I
have to copy the link from Thunderbird to Opera to get to the website. What
happened??


--
Everybody's gotta price. No shame, it's all a game, just ask it
nice..........

Bit Twister wrote:
> On Sat, 27 Sep 2008 23:19:31 -0400, Jim Beard wrote:
>
>> Bit Twister wrote:
>>> I suggest you use the --wget option. Some sites only allow one
>>> connection at a time from the same ip. If curl runs instead of wget,
>>> you could miss some files.
>> The above does not seem to work properly when the repository
>> is accessed by rsync. For example, the following 3 lines appear
>> in a single burst, requiring a fraction of a second.
>>
>> rsync://carroll.cac.psu.edu/mandrakelinux/official/2008.1/x86_64/media/non-free/backports/media_info/synthesis.hdlist.cz
>> ...retrieving failed: rsync failed: exited with 10
>>
>> problem reading synthesis file of medium "Non-free Backports"
>>
>> On the other hand, the following works fine for me.
>> urpmi --wget --auto-update --auto-select --auto
>
> Heheh, carroll is my favorite site for the single connect/ip failure.
> I have a script which reads /etc/urpmi/urpmi.cfg and pull those directory's
> package names, dates, size into an .idx file. Much faster greping *.idx
> than using a urpm tool to find a package name. :-)
>
> I had several retrieving failed messages, so I put a while loop waiting for
> the last wget connection to drop before fetching the next directory.
>
> Playing with cooker loads, from the command line, gives me the idea
> that you might have drop outs depending on the size of the packages
> being installed/removed.
>
> I can guess your backports rsync was caused by the previous directory
> having something or not downloaded. Unsure which.
>
> I have taken to running the
> urpmi --wget --auto-update --auto-select --auto
> until I get all packages are up to date.
>
> PS: carroll has been down since about 4pm Saturday. Guessing maintenance.

If the site has been completely, how is it the following _appeared_
to run as normal and notified me that all packages were up to date?
(This was expected, as I had updated the night before, and I had seen
nothing in Mandriva Security forum that made me expect something new.)

urpmi --wget --auto-update --auto-select --auto

If the above is failing to work, as has been said to happen, how
is it that I have had no problems? And I do monitor Mandriva
[Security Announce] to ensure that kernels and other things that
might not be automatically updated do not get past me unnoticed.

To me, things are a little murky at the moment. Perhaps the route to
Carroll makes a difference? I am in Northern Virginia, so fairly
close.

I do have problems with slow connections to Mandriva sites in France,
with some uncertainty about which mode (ftp, rsync) or site (Paris,
Orleans, other) will work best at any given time. I had simply put
that down to international routing plus normal problems with French
infrastructure and service, but maybe the problem is at my end?

Cheers!

jim b.

--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

Dallas wrote:
> I have recently upgraded to version 2008.1 and have noticed a lost function
> in the upgrade. I'm running Thunderbird as my mail client, and Opera 9.52
> as my browser. In my browser, I can click an email address, and thunderbird
> will open, but I can't click a link in Thunderbird and get Opera to open. I
> have to copy the link from Thunderbird to Opera to get to the website. What
> happened??

If using KDE, have you set your default browser in KDE Control Center
--> KDE Components -- > Component Choose ?

Cheers!

jim b.


--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

On Sun, 28 Sep 2008 07:46:20 -0400, Jim Beard wrote:
>
> If the site has been completely, how is it the following _appeared_
> to run as normal and notified me that all packages were up to date?

In the past I have seen "urpmi.update -a" fail on some connection and
yet return a good status. I changed my script to use each mirror
instead of -a and would a failure status on failed mirrors.

No idea if your rsync connector errors are hidden or if just rsync is
the only thing up.

Augh frap, now it is starting to work just as I try to get proof. :-(


When I had my urpmi failures I would click these links and they both
would fail.
ftp://carroll.cac.psu.edu/pub/linux/...mandrivalinux/
http://carroll.cac.psu.edu/pub/linux...inux/official/

It is intermittent though, and directory contents are empty when I do
get in.


> (This was expected, as I had updated the night before, and I had seen
> nothing in Mandriva Security forum that made me expect something new.)
>
> urpmi --wget --auto-update --auto-select --auto
>
> If the above is failing to work, as has been said to happen, how
> is it that I have had no problems? And I do monitor Mandriva
> [Security Announce] to ensure that kernels and other things that
> might not be automatically updated do not get past me unnoticed.

no idea about 64 bit, 32 bit has.
2008-09-26 MDVSA-2008:206 mozilla-thunderbird
2008-09-25 MDVSA-2008:205 mozilla-firefox
2008-09-24 MDVSA-2008:204 blender




> To me, things are a little murky at the moment. Perhaps the route to
> Carroll makes a difference? I am in Northern Virginia, so fairly
> close.

No, getting a solid fail on
http://carroll.cac.psu.edu/pub/linux...inux/official/

and empty contents for
ftp://carroll.cac.psu.edu/pub/linux/...mandrivalinux/

Bit Twister wrote:
> On Sun, 28 Sep 2008 07:46:20 -0400, Jim Beard wrote:
>> If the site has been completely, how is it the following _appeared_
>> to run as normal and notified me that all packages were up to date?
>
> In the past I have seen "urpmi.update -a" fail on some connection and
> yet return a good status. I changed my script to use each mirror
> instead of -a and would a failure status on failed mirrors.
>
> No idea if your rsync connector errors are hidden or if just rsync is
> the only thing up.
>
> Augh frap, now it is starting to work just as I try to get proof. :-(
>
>
> When I had my urpmi failures I would click these links and they both
> would fail.
> ftp://carroll.cac.psu.edu/pub/linux/...mandrivalinux/
> http://carroll.cac.psu.edu/pub/linux...inux/official/
>
> It is intermittent though, and directory contents are empty when I do
> get in.
>
>
>> (This was expected, as I had updated the night before, and I had seen
>> nothing in Mandriva Security forum that made me expect something new.)
>>
>> urpmi --wget --auto-update --auto-select --auto
>>
>> If the above is failing to work, as has been said to happen, how
>> is it that I have had no problems? And I do monitor Mandriva
>> [Security Announce] to ensure that kernels and other things that
>> might not be automatically updated do not get past me unnoticed.
>
> no idea about 64 bit, 32 bit has.
> 2008-09-26 MDVSA-2008:206 mozilla-thunderbird
> 2008-09-25 MDVSA-2008:205 mozilla-firefox
> 2008-09-24 MDVSA-2008:204 blender
>
>
>
>
>> To me, things are a little murky at the moment. Perhaps the route to
>> Carroll makes a difference? I am in Northern Virginia, so fairly
>> close.
>
> No, getting a solid fail on
> http://carroll.cac.psu.edu/pub/linux...inux/official/
>
> and empty contents for
> ftp://carroll.cac.psu.edu/pub/linux/...mandrivalinux/

Thunderbird and Firefox both updated to 2.0.0.17-1.1, 32-bit and
64-bit. I do not have blender installed.

I follow the logic of what you are saying, and that would explain
what I am seeing. Still, I generally have gotten updates as
expected, though sometimes a day later than expected. This is so
much better than service from secsup.org that I have been quite
happy with it, and never dreamed that problems lurked.

And now both the following are failing, with identical error
messages. faaaarrrggh.

urpmi --wget --auto-update --auto-select --auto
urpmi --wget --update --auto-update --auto

Cheers!

jim b.

--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

On Sun, 28 Sep 2008 08:57:45 -0400, Jim Beard wrote:
>
> I follow the logic of what you are saying, and that would explain
> what I am seeing. Still, I generally have gotten updates as
> expected, though sometimes a day later than expected.

Yep, I used carroll because it seemed to be feeding other USA mirrors.

> This is so
> much better than service from secsup.org that I have been quite
> happy with it, and never dreamed that problems lurked.

Heheh, I liked the problems on carroll in the 9/10 release days.
Everyone would pick another mirror and it was always pretty spry for
me.

> And now both the following are failing, with identical error
> messages. faaaarrrggh.

Yes, glad to hear it. I was wondering if urpmi was hiding connection
failures.

You might consider gatech.edu

Here run these two commands

host ftp://carroll.cac.psu.edu
host ftp.gtlib.gatech.edu


--
http://uncyclopedia.org/wiki/Microsoft

Dallas skrev:
> I have recently upgraded to version 2008.1 and have noticed a lost function
> in the upgrade. I'm running Thunderbird as my mail client, and Opera 9.52
> as my browser. In my browser, I can click an email address, and thunderbird
> will open, but I can't click a link in Thunderbird and get Opera to open. I
> have to copy the link from Thunderbird to Opera to get to the website. What
> happened??
>
>
Hi, I am using the Norwegian version of Mandriva, but you should find
Opera going here, Menu->Controlpanel->Componentchooser and choose
Netbrowser, then you can browse to find Opera. (Hope I have translated
properly here)

All the best

Roger

--
Mandriva Linux 2008.1 pwp (Linuxcounter #432950)

MSI K8TM, 1,0 Gb Memory, AMD Sempron 3000+ CPU
ATI Sapphire Radeon X1600
__________________________________________________
Be against others the way
you want to be treated yourself

Bit Twister wrote:
> On Sun, 28 Sep 2008 08:57:45 -0400, Jim Beard wrote:
>> I follow the logic of what you are saying, and that would explain
>> what I am seeing. Still, I generally have gotten updates as
>> expected, though sometimes a day later than expected.
>
> Yep, I used carroll because it seemed to be feeding other USA mirrors.
>
>> This is so
>> much better than service from secsup.org that I have been quite
>> happy with it, and never dreamed that problems lurked.
>
> Heheh, I liked the problems on carroll in the 9/10 release days.
> Everyone would pick another mirror and it was always pretty spry for
> me.
>
>> And now both the following are failing, with identical error
>> messages. faaaarrrggh.
>
> Yes, glad to hear it. I was wondering if urpmi was hiding connection
> failures.
>
> You might consider gatech.edu
>
> Here run these two commands
>
> host ftp://carroll.cac.psu.edu
> host ftp.gtlib.gatech.edu

Back in the days of 8.1 through maybe 10 or LE2005 (? was that the
name?), Georgia Tech was sluggish and erratic, in my experience.
Over the last year, though, it has become by backup of choice, when
carroll was not available or overloaded. I do not remember the
gtlib part though. Is this a new alias? Or have they added a
separate machine/server name?

I may have to take a look at that the next time I am in the mood to
tinker with my sources. I am currently unhappy with slowness of
my connection to plf in France, and may need to change servers there.

Recommendations?

Cheers!

jim b.

--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.

On Sun, 28 Sep 2008 20:34:41 -0400, Jim Beard wrote:
> Bit Twister wrote:
>> Here run these two commands
>>
>> host ftp://carroll.cac.psu.edu
>> host ftp.gtlib.gatech.edu
>
> Back in the days of 8.1 through maybe 10 or LE2005 (? was that the
> name?), Georgia Tech was sluggish and erratic, in my experience.

Yep, same here.

> Over the last year, though, it has become by backup of choice, when
> carroll was not available or overloaded. I do not remember the
> gtlib part though. Is this a new alias? Or have they added a
> separate machine/server name?

Try the host commands :-)


> I may have to take a look at that the next time I am in the mood to
> tinker with my sources.

carroll still down as I type. gatech could maintain full throughput on
my 15/2 fios connection when downloading an iso.

> I am currently unhappy with slowness of
> my connection to plf in France, and may need to change servers there.
>
> Recommendations?

Have not tested any in france. On slow days, you might check
http://www.internetpulse.net/
before assuming the site is slow.

Might not hurt to run
traceroute -I each_mirror_here

On Sun, 28 Sep 2008 22:27:07 -0400, Bit Twister wrote:

> carroll still down as I type. gatech could maintain full throughput on
> my 15/2 fios connection when downloading an iso.

I've switched back to mirrors.secsup.org for now. I know, from resolving
prior problems, that carroll gets it's feed directly from Mandriva, and
mirrors.secsup.org gets it's feed from carroll.

secsup tends to be a few hours behind carroll in getting new updates, but
transfers tend to be faster, and "too many users" messages are much rarer
(never seen one from secsup).

With the recent problems at carroll, I think I've found a bug in urpmi.
If the ftp connection fails, even with the "--resume" and "--noclean" options
specified, the previously downloaded partial rpm will be deleted. That's
a royal pain with dialup. I ended up using wget to download the rpm for
blender, to prevent the download from restarting at the beginning everytime
my dialup connection dropped.

That brings up another topic. My intel-536EP winmodem seems to be slowly
dying. I used to normally get connections at 46600, but for the last month
or so, am only getting 28800 (if I'm lucky). When it connects at even lower
speeds, the connection usually drops, in under 20 minutes. Using a friends
laptop, I was able to connect consistently at 45.2, using the same phone line,
so it isn't a phone line problem. The problem with connection speeds seems
to coincide with the problem I posted about almost a month ago, where I lost
two file systems. At this point, I'm not sure if it's a hardware problem, or
a kernel driver problem.

Regards, Dave Hodgins

--
Change nomail.afraid.org to ody.ca to reply by email.
(nomail.afraid.org has been set up specifically for
use in usenet. Feel free to use it yourself.)

Bit Twister wrote:

> On Sat, 27 Sep 2008 22:47:41 +1000, Rob wrote:
>> I have recently upgraded to version 2008.1 and have noticed an apparent
>> difference between this version and 2007.1 which was the previous version
>> I used.
>>
>> With the earlier version, if the automatic mandriva update applet showed
>> updates were available, I could then use the 'urpmi --auto-update'
>> command
>> and it would find and install the updates. With 2008.1 if the applet
>> indicates there are updates and I then run the manual update *no* updates
>> are found. However if I then use the applet to perform the installation
>> there are indeed updates available.
>
>
> If you are going to run from the command line, use something like
> urpmi --wget --update --auto-update --auto
>
> That will verify your rpm lists match the mirror's list.
> If different, download the new lists used by auto-update.
> Just running with --auto-update only checks your rpm list against your
> install.
>
> I suggest you use the --wget option. Some sites only allow one
> connection at a time from the same ip. If curl runs instead of wget,
> you could miss some files.

Thanks for these suggestions - they obviously makes all the difference! Now
I need to understand the whys. I'd always understood that:

urpmi --auto-update was equivalent to urpmi.update -a && urpmi --auto-select
so that the update bit was actually carried out by that. I was certain
from observation that the --auto-update version *did* actually carry out
the updates to local hdlist files in (at least) 2007.0 & 2007.1 versions of
Mandriva Linux.

I just looked in /var/lib/urpmi and noticed that the file set looks
different from what it was in earlier versions I'd seen so that may well be
the sticking point. Also I noticed that when I use -v (verbose) option
that two local files are compared whereas previously only one local file
was compared to the remote one - seemingly a different mechanism than
before...

Anyway, at least now I know what actually *does* work from the command line.

Thanks again

Rob.

David W. Hodgins wrote:
> On Sun, 28 Sep 2008 22:27:07 -0400, Bit Twister wrote:
>
>> carroll still down as I type. gatech could maintain full throughput on
>> my 15/2 fios connection when downloading an iso.
>
> I've switched back to mirrors.secsup.org for now. I know, from resolving
> prior problems, that carroll gets it's feed directly from Mandriva, and
> mirrors.secsup.org gets it's feed from carroll.

> That brings up another topic. My intel-536EP winmodem seems to be slowly
> dying. I used to normally get connections at 46600, but for the last month
> or so, am only getting 28800 (if I'm lucky). When it connects at even lower
> speeds, the connection usually drops, in under 20 minutes. Using a friends
> laptop, I was able to connect consistently at 45.2, using the same phone line,
> so it isn't a phone line problem. The problem with connection speeds seems
> to coincide with the problem I posted about almost a month ago, where I lost
> two file systems. At this point, I'm not sure if it's a hardware problem, or
> a kernel driver problem.

Have you checked your modem settings and the settings in
configuration files under /etc? It has been some years since I was
on dial-up, but I found that tinkering with the settings was
critical, and that every few months a power surge or some such would
cause the modem to revert to defaults. Settings in the Linux files
could also be restoring things to less-than-optimal defaults.

Cheers!

jim b.

--
UNIX is not user unfriendly; it merely
expects users to be computer-friendly.