Apt has strange network errors when updating indexes.
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
apt (Ubuntu) |
Invalid
|
Medium
|
Unassigned | ||
Bug Description
Binary package hint: apt
If I update the apt indices either in Synaptic Package Manager, Update Manager or from the command line, it fails to download a random selection of indexes (from 0 to 3, usually). The errors look like this:
W: Failed to fetch http://
W: Failed to fetch http://
E: Some index files failed to download, they have been ignored, or old ones used instead.
It is usually about two that fail, though sometimes more, and sometimes none. I just have a wireless router attached to a cable modem, standard home broadband. If I run 'update' a couple times, it will get everything. Apt has no problem downloading or installing packages. I can run 'ping' against these servers with zero packet loss.
If I do a single update with errors and then try to install packages, sometimes it warns me that packages are unsigned.
Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't an upgrade problem, I backed up my personal files, and installed 10.04 from scratch. The problem persisted with the default Ubuntu 10.04 install.
My laptop is a Compaq CQ60 215dx.
ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: apt 0.7.25.3ubuntu7
ProcVersionSign
Uname: Linux 2.6.32-21-generic x86_64
Architecture: amd64
Date: Mon May 3 19:33:15 2010
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
ProcEnviron:
LANG=en_US.utf8
SHELL=/bin/bash
SourcePackage: apt
Stephen A. Goss (postfuturist) wrote : | #1 |
AlphaZeta (kerryw) wrote : | #2 |
diddy1234 (richarddidd) wrote : | #3 |
I can also confirm that here in the UK the problem exists as well.
I also checked each site link in a web browser and I can get to these web sites.
I also checked my etc/hosts file and this was correctly populated.
output from 'apt-get update' :-
Hit http://
Ign http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Ign http://
Hit http://
Hit http://
Hit http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Err http://
iceman21ad (cmdunionmillwright) wrote : | #4 |
looks like the signatures are not downloading
chris@chris-
[sudo] password for chris:
Get:1 http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
That Bum (jzachariou) wrote : | #5 |
This affects me too.
See:
http://
http://
Changed in apt (Ubuntu): | |
status: | New → Confirmed |
YeagDeeDs (yeagdeeds) wrote : | #6 |
It appears to work correctly when I am not using a third party repository. But if I add any others besides the defaults it throws the Wicked error. I am using Ubuntu 10.04 64bit.
guymac (guymac-gmail) wrote : | #7 |
I'm getting errors for security.ubuntu.com now, as well as packages.
W: GPG error: http://
W: Failed to fetch http://
Dan (dorjedan) wrote : | #8 |
I get the same errors... security.
manishbhoola (manishbhoola) wrote : | #9 |
I am using the India servers for update. I too get the following errors consistently for the same repos.
W: Failed to fetch http://
W: Failed to fetch http://
I have the following repos active
deb http://
deb-src http://
deb http://
deb-src http://
deb http://
deb-src http://
deb http://
deb http://
Sven Heinicke (svenmh) wrote : | #10 |
I connect at home (PA, USA) via Comcast and I have this problem, but when I connect to work though VPN all works will.
jeff (jeff.r) wrote : | #11 |
Same situation across all 3 systems in our house. 2 upgraded from 9.10 and one a clean install of Lucid
Get:1 http://
Ign http://
Hit http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Get:2 http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
Err http://
Something wicked happened resolving 'packages.m...
That Bum (jzachariou) wrote : | #12 |
Put an update on this thread: http://
To summarize what I said so far, I have bad DSL line quality (uncanceled echo) that's making my router flip out and randomly drop the connection. This also appears to be causing this problem with apt-get update, through there is no packet loss while the router near as I can tell, and latency seems to be OK.
How many people here have DSL?
If you have a router/modem thing, look in its config, however you do it with your model. Do you have uncanceled echo?
Trying to find a common link here.
Ivan Zorin (iaz) wrote : | #13 |
I can confirm this bug - in my case apt can't get access to security.
> sudo apt-get update
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Josh Brown (joshbrown) wrote : | #14 |
I believe this Ubuntu Forums thread is related: http://
YeagDeeDs (yeagdeeds) wrote : | #15 |
I tried installing a new copy of kubuntu and installing all updates. The problem is still present and erratic(other than only appearing after adding any repository). I have tried both Kubuntu and Ubuntu with the same results.
Stephen A. Goss (postfuturist) wrote : | #16 |
It's somehow related to the internet connection, as I only get the error on my home internet connection (wireless 802.11g router with WPA2 connected to Comcast cable modem), but not when I'm at work (wireless 802.11g router with WEP 128 bit connected to local ISP).
Dan (dorjedan) wrote : | #17 |
It may be true that it is somehow related to the internet connection... But it is an issue that is unique to Lucid since I use the same connection that I always have, and never experienced this error in Karmic.
That Bum (jzachariou) wrote : | #18 |
I've found a solution. A real one.
It seems that it's a problem with Network Manager. People who have routers with gateways will make Network Manager incorrectly set the DNS to the router's internal address. This only works if everything is set to automatic. If you set the DNS manually in Network Manager, this can be fixed completely, provided you know the DNS's address, presumably from going on the router's config page and looking for it there if you haven't seen it.
Credit to newb85 on post #17 on http://
manishbhoola (manishbhoola) wrote : Re: [Bug 574886] Re: Apt has strange network errors when updating indexes. | #19 |
Bingo - solved the problem for me !! Thanks for posting.
On Wed, May 19, 2010 at 7:33 AM, That Bum <email address hidden> wrote:
> I've found a solution. A real one.
>
> It seems that it's a problem with Network Manager. People who have
> routers with gateways will make Network Manager incorrectly set the DNS
> to the router's internal address. This only works if everything is set
> to automatic. If you set the DNS manually in Network Manager, this can
> be fixed completely, provided you know the DNS's address, presumably
> from going on the router's config page and looking for it there if you
> haven't seen it.
>
> Credit to newb85 on post #17 on
> http://
>
> --
> Apt has strange network errors when updating indexes.
> https:/
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “apt” package in Ubuntu: Confirmed
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random selection of
> indexes (from 0 to 3, usually). The errors look like this:
>
> W: Failed to fetch
> http://
> (-5 - No address associated with hostname)
>
> W: Failed to fetch http://
> (-5 - No address associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old ones
> used instead.
>
> It is usually about two that fail, though sometimes more, and sometimes
> none. I just have a wireless router attached to a cable modem, standard home
> broadband. If I run 'update' a couple times, it will get everything. Apt has
> no problem downloading or installing packages. I can run 'ping' against
> these servers with zero packet loss.
>
> If I do a single update with errors and then try to install packages,
> sometimes it warns me that packages are unsigned.
>
> Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't an
> upgrade problem, I backed up my personal files, and installed 10.04 from
> scratch. The problem persisted with the default Ubuntu 10.04 install.
>
> My laptop is a Compaq CQ60 215dx.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 10.04
> Package: apt 0.7.25.3ubuntu7
> ProcVersionSign
> Uname: Linux 2.6.32-21-generic x86_64
> Architecture: amd64
> Date: Mon May 3 19:33:15 2010
> InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
> ProcEnviron:
> LANG=en_US.utf8
> SHELL=/bin/bash
> SourcePackage: apt
>
> To unsubscribe from this bug, go to:
> https:/
>
Craig100 (craig-incrediblyuseful) wrote : | #20 |
Didn't work for me:(
Running Lucid x64
When I reload repositories from the UK or Main server in Synaptic I get the following error message:-
GPG error: http://
Some index files failed to download, they have been ignored, or old ones used instead.
My sources list is:-
# deb cdrom:[Ubuntu 10.04 LTS _Lucid Lynx_ - Release amd64 (20100429)]/ lucid main restricted
# See http://
# newer versions of the distribution.
deb http://
deb-src http://
## Major bug fix updates produced after the final release of the
## distribution.
deb http://
deb-src http://
## N.B. software from this repository is ENTIRELY UNSUPPORTED by the Ubuntu
## team. Also, please note that software in universe WILL NOT receive any
## review or updates from the Ubuntu security team.
deb http://
deb-src http://
deb http://
deb-src http://
## N.B. software from this repository is ENTIRELY UNSUPPORTED by the Ubuntu
## team, and may not be under a free licence. Please satisfy yourself as to
## your rights to use the software. Also, please note that software in
## multiverse WILL NOT receive any review or updates from the Ubuntu
## security team.
deb http://
deb-src http://
deb http://
deb-src http://
## Uncomment the following two lines to add software from the 'backports'
## repository.
## N.B. software from this repository may not have been tested as
## extensively as that contained in the main release, although it includes
## newer versions of some applications which may provide useful features.
## Also, please note that software in backports WILL NOT receive any review
## or updates from the Ubuntu security team.
# deb http://
# deb-src http://
## Uncomment the following two lines to add software from Canonical's
## 'partner' repository.
## This software ...
Craig100 (craig-incrediblyuseful) wrote : | #21 |
Further top my last, I'm also having issues with Funambol synching in Thunderbird 3. Basically, TB3 crashes as soon as Funambol accesses the network. Could be another clue as this started at the same time Synaptic started having issues.
dazzlin (dazzlindonna) wrote : | #22 |
I can confirm this problem as well. It only began happening to me when I upgraded to Mint 9. I've tried various suggested solutions as outlined in this thread http://
That Bum (jzachariou) wrote : | #23 |
Craig100,
You are using OpenDNS, yes? Your secondary DNS IP is incorrect, it's not 208.68.220.220, it's 208.67.220.220.
See OpenDNS's homepage, http://
faortiz (fortiz-usa) wrote : | #24 |
I have tried all the suggestions listed here in this thread, and I still see some of the important repositories being ignored:
Ign cdrom://Ubuntu 9.10 _Karmic Koala_ - Release i386 (20091028.5)/ karmic/main Translation-en_US
Ign cdrom://Ubuntu 9.10 _Karmic Koala_ - Release i386 (20091028.5)/ karmic/restricted Translation-en_US
Hit http://
Ign http://
Get:1 http://
Hit http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Hit http://
Ign http://
Ign http://
Ign http://
Ign http://
Hit http://
Hit http://
Ign http://
Get:2 http://
Hit http://
Ign http://
Hit http://
Hit http://
Get:3 http://
Hit http://
Get:4 http://
Hit http://
Hit http://
Hit http://
Hit http://
Hit http://
That Bum (jzachariou) wrote : | #25 |
faortiz,
This sounds like a totally different issue there, ignoring repositories. In this bug, apt doesn't ignore them but has difficulty resolving them.
Also, it seems you've dist-upgraded from Karmic to Lucid...this is most likely a factor. Try a fresh install of Lucid.
himanshu.p.singh (himanshu-p-singh) wrote : | #26 |
Everything was fine for me when I did a from_scratch install of 10.04 till I tried to install a package called gorm.app and on including any of its debian mirrors (http://
The errors seem to be of two types:
First Type:
if I use any of the mirror urls like: www.anheng.
and have this line in my sources.list: deb http://
Then I see the following error message:
Get:3 http://
Ign http://
Err http://
Something wicked happened resolving 'archive.
W: Failed to fetch http://
Second Type:
if I use any of the mirror urls like: mirror.
and have this line in my sources.list: deb http://
Then I see the following error message:
Err http://
Something wicked happened resolving 'mirror.
Ign http://
Get:2 http://
Get:3 http://
Fetched 6,551kB in 3min 49s (28.5kB/s)
W: Failed to fetch http://
And if I remove this " deb http://
himanshu.p.singh (himanshu-p-singh) wrote : | #27 |
BTW, I am using http://
himanshu.p.singh (himanshu-p-singh) wrote : | #28 |
OK.. found a workaround.. but not a really good one...
turns out that this error was related to my using a Router... and when I disconnected the router and hooked my comp straight to the modem, the problem of "-5 - No address associated with hostname" was resolved. Probably others can try the same too to see if it helps...
On looking further, I found that someone else has stumbled upon this too... (though the workaround proposed there looks to be more complete.. but I have not yet tried it).. see post #17 below:
http://
Happy debugging :) .. finally can get some sleep
faortiz (fortiz-usa) wrote : | #29 |
I'm sorry: I guess I mis read what was being said. :( I was able to resolve my issue by re-writing the repository list. Thank you for the the suggestion, though. :)
Winston Ewert (winstonewert) wrote : | #30 |
This problem went away for me when I replaced my router (which was failing in other ways as well, I had to constantly restart it to keep it running.)
Ali Servet Donmez (exalted) wrote : | #31 |
Heads up people, won't we have an official word on this one? I mean what's good for an LTS release if you can't even update your packages?
Also, I believe this is not an amd64 only issue, since I'm having same messages also in a fresh 10.04 x86 server install.
Regards,
Jesse R. Taylor (jrtayloriv) wrote : | #32 |
I am experiencing the same issue. As Dan wrote in #17, regardless of whether this is caused by "using a router", and can be "fixed" by plugging directly into the modem etc. it is still a major bug that needs to be fixed. Before upgrading to 10.04, everything worked fine with my current network setup, and now it doesn't. Replacing routers or not using them at all is not a solution.
Jesse R. Taylor (jrtayloriv) wrote : | #33 |
The following Debian bug report might be of interest: http://
Alan Weeks (aweeks1) wrote : | #34 |
I too, have problems with the "something wicked" bug. I cannot now update/upgrade browse or receive email;
I upgraded 8.04 to 10.04. As it turns out, a mistake!
I have two disappointments---
1. I believed that LTS would "work out of the box" & was stable
2. Some of the intended help above are a bit brief for a Ubuntu USER, as distinct from a techie.
I tied the one about the visudo & sudoers, but with no success. I cannot write to sudoers, but keep getting extra files named .temp or .temp.1 etc.
I use a router (#17 above) but even without router problem still exists.
Changing the sites for updates from Australi to Main makes no difference.
I would appreciate any kind soul writing either a fix or a workaround such that even a relative newcomer such as I am can understand it.
(This means telling me when to use SUDO, when to REBOOT, where any fiile is located and so on!
Some of us have NOT used GNU/Linux for years!)
Jason DePriest (jrdepriest) wrote : | #35 |
First, let me say that using the suggestion in #18 about hard coding the DNS fixed the problem for me.
I have a laptop with a 32-bit AMD Sempron processor that I wiped and did a full, fresh install of Ubuntu Lucid Lynx.
I have added some 3rd party repositories in my /etc/apt/sources.d/ directory.
I am on a home wireless network 802.11g with WPA2 + PSK. My router is a 2Wire Residential Gateway device that acts as DSL modem, wireless router, firewall, and is also where our AT&T U-Verse television comes in the house.
I do not have a proxy server.
By default, Ubuntu was using the RG itself as the DNS server which is supposed to work just fine. It does work in almost all cases. It doesn't work for apt-get for some reason. I was able to log on to my RG's web interface and get the DNS servers that it was using. I then setup my network connection to use those instead.
That fixed it.
But, why is this necessary?
Also, LTS? I think you've found a new scenario to add to your QA testing.
MichalM (jafa81) wrote : | #36 |
Hi Jason,
I have the same scenario - AT&T U-verse and issues with APT.
Issue disappeared when I manually set DNS to google's free DNS service (
8.8.8.8 and 8.8.4.4 , pretty fast).
However, I tried latest Alpha of Ubuntu 10.10 and it DOES NOT have this
issue anymore!
What I have noticed earlier, AT&T's DNS servers are very slow. Typical query
takes somewhere around 100ms. But every once in a while (about every 20th
query in my case) takes about 500ms.
This is very annoying even when surfing the net. average www page initiates
dozens of DNS queries, which means I hit the issue with slow response pretty
much every time I open a web page.
I think APT doesn't handle well slow DNS responses and times out internally.
Speaking of 2WIRE gateway, have you found a way how to manually change DNS
directly on the box? I'd like to prefer this option over setting every
laptop manually.
Mike
On Tue, Jul 13, 2010 at 12:37 PM, Jason DePriest
<email address hidden>wrote:
> First, let me say that using the suggestion in #18 about hard coding the
> DNS fixed the problem for me.
>
> I have a laptop with a 32-bit AMD Sempron processor that I wiped and did
> a full, fresh install of Ubuntu Lucid Lynx.
>
> I have added some 3rd party repositories in my /etc/apt/sources.d/
> directory.
>
> I am on a home wireless network 802.11g with WPA2 + PSK. My router is a
> 2Wire Residential Gateway device that acts as DSL modem, wireless
> router, firewall, and is also where our AT&T U-Verse television comes in
> the house.
>
> I do not have a proxy server.
>
> By default, Ubuntu was using the RG itself as the DNS server which is
> supposed to work just fine. It does work in almost all cases. It
> doesn't work for apt-get for some reason. I was able to log on to my
> RG's web interface and get the DNS servers that it was using. I then
> setup my network connection to use those instead.
>
> That fixed it.
>
> But, why is this necessary?
>
> Also, LTS? I think you've found a new scenario to add to your QA
> testing.
>
> --
> Apt has strange network errors when updating indexes.
> https:/
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “apt” package in Ubuntu: Confirmed
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random selection of
> indexes (from 0 to 3, usually). The errors look like this:
>
> W: Failed to fetch
> http://
> (-5 - No address associated with hostname)
>
> W: Failed to fetch http://
> (-5 - No address associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old ones
> used instead.
>
> It is usually about two that fail, though sometimes more, and sometimes
> none. I just have a wireless router attached to a cable modem, standard home
> broadband. If I run 'upd...
Jason DePriest (jrdepriest) wrote : | #37 |
I have tried to find a way to manually set the DNS and I can't find it.
There used to be a special advanced config area you could get to by
doing to /mdc (or something like that) that let you change more
settings, but it wasn't available there either.
Since my RG has been updated (automatically) from the 5.x series of
code to the 6.x series, even that is no longer available.
There may be other hidden settings, but I'm sure that won't be
supported by AT&T.
You'd have to host your own router on the other side of the 2Wire RG
and have it do the DNS the way you want it. That was too much trouble
for me, so I just use that AT&T gives me.
I had no idea that there were DNS latency problems like that. I've
never done any testing.
MichalM (jafa81) wrote : | #38 |
Jason,
I know /mdc disappeared for a while but now it's back. My gateway
runs version 6.1.9.24-enh.tm and I can get to /mdc. But it's more like a
detailed list of setting than actual configuration page.
If you want to test your AT&T's DNS servers you can use this tool:
http://
It's free but Windoze only :-(
Unfortunately I did not find any Linux alternative as simple and nice as
this one.
Mike
On Tue, Jul 13, 2010 at 1:41 PM, Jason DePriest
<email address hidden>wrote:
> I have tried to find a way to manually set the DNS and I can't find it.
>
> There used to be a special advanced config area you could get to by
> doing to /mdc (or something like that) that let you change more
> settings, but it wasn't available there either.
>
> Since my RG has been updated (automatically) from the 5.x series of
> code to the 6.x series, even that is no longer available.
>
> There may be other hidden settings, but I'm sure that won't be
> supported by AT&T.
>
> You'd have to host your own router on the other side of the 2Wire RG
> and have it do the DNS the way you want it. That was too much trouble
> for me, so I just use that AT&T gives me.
>
> I had no idea that there were DNS latency problems like that. I've
> never done any testing.
>
> --
> Apt has strange network errors when updating indexes.
> https:/
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “apt” package in Ubuntu: Confirmed
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random selection of
> indexes (from 0 to 3, usually). The errors look like this:
>
> W: Failed to fetch
> http://
> (-5 - No address associated with hostname)
>
> W: Failed to fetch http://
> (-5 - No address associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old ones
> used instead.
>
> It is usually about two that fail, though sometimes more, and sometimes
> none. I just have a wireless router attached to a cable modem, standard home
> broadband. If I run 'update' a couple times, it will get everything. Apt has
> no problem downloading or installing packages. I can run 'ping' against
> these servers with zero packet loss.
>
> If I do a single update with errors and then try to install packages,
> sometimes it warns me that packages are unsigned.
>
> Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't an
> upgrade problem, I backed up my personal files, and installed 10.04 from
> scratch. The problem persisted with the default Ubuntu 10.04 install.
>
> My laptop is a Compaq CQ60 215dx.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 10.04
> Package: apt 0.7.25.3ubuntu7
> ProcVersionSign
> Uname: Linux 2.6.32-21-generic x86_64
>...
pyutaros (pyutaros) wrote : | #39 |
Changing DNS to Google (8.8.8.8) worked for me. AT&T Uverse customer with 2Wire modem.
That Bum (jzachariou) wrote : | #40 |
Indeed, AT&T's DNS is slow. AT&T is my ISP as well, and I also have a 2wire router, however, I have DSL and not U-verse, which is cable.
A DNS tester for Linux would be here http://
(Yes, the screenies show OS X, that's because it's multi-platform and they all look the same.)
Paul van Genderen (paulvg) wrote : | #41 |
I'm not sure if changing this to ¨address only¨ will correctly configure the default route, which I have no control over. Also, /etc/resolv.conf looks like this:
# Generated by NetworkManager
nameserver 4.2.2.1
nameserver 4.2.2.2
These are not the nameservers of my ISP. The correct resolv.conf seems to sit in /etc/ppp/.
This bug does not appear in aptitude.
Jordan Farrell (feralbytes) wrote : | #42 |
I was affected by this bug; but I am in a horrible location for Internet and behind a proxy server. But I fixed my bug by deleting all of the none default hidden folders in my root directory. I used gksu nautilus in order to have the correct permissions. I then rebooted my machine. Once I logged back in my machine was behaving normally again. Hope that helps.
Heimen Stoffels (vistaus) wrote : | #43 |
It still happens here. Can somebody please fix this bug?
Phillip Castaneda (pcastaneda) wrote : | #44 |
"Changing DNS to Google (8.8.8.8) worked for me. AT&T Uverse customer with 2Wire modem."
--how did you do this?
@MichaL I'm having the same problem with new firmware. Tech support (via chat) said this was a bug they were urgently working on, but... still wondering if someone figured a way to change it
Thomas Rolland Butler (fatchickassassin) wrote : | #45 |
Phillip go into your router (Browser address bar http://
MichalM (jafa81) wrote : Re: [Bug 574886] Re: Apt has strange network errors when updating indexes. | #46 |
Phillip,
Unfortunately 2WIRE routers (that come with Uverse) don't allow you to set
custom DNS servers. I had to change DNS on every host computer :-(.
MichaL
On Sun, Aug 15, 2010 at 9:10 PM, Phillip Castaneda <
<email address hidden>> wrote:
> "Changing DNS to Google (8.8.8.8) worked for me. AT&T Uverse customer with
> 2Wire modem."
> --how did you do this?
> @MichaL I'm having the same problem with new firmware. Tech support (via
> chat) said this was a bug they were urgently working on, but... still
> wondering if someone figured a way to change it
>
> --
> Apt has strange network errors when updating indexes.
> https:/
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “apt” package in Ubuntu: Confirmed
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random selection of
> indexes (from 0 to 3, usually). The errors look like this:
>
> W: Failed to fetch
> http://
> (-5 - No address associated with hostname)
>
> W: Failed to fetch http://
> (-5 - No address associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old ones
> used instead.
>
> It is usually about two that fail, though sometimes more, and sometimes
> none. I just have a wireless router attached to a cable modem, standard home
> broadband. If I run 'update' a couple times, it will get everything. Apt has
> no problem downloading or installing packages. I can run 'ping' against
> these servers with zero packet loss.
>
> If I do a single update with errors and then try to install packages,
> sometimes it warns me that packages are unsigned.
>
> Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't an
> upgrade problem, I backed up my personal files, and installed 10.04 from
> scratch. The problem persisted with the default Ubuntu 10.04 install.
>
> My laptop is a Compaq CQ60 215dx.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 10.04
> Package: apt 0.7.25.3ubuntu7
> ProcVersionSign
> Uname: Linux 2.6.32-21-generic x86_64
> Architecture: amd64
> Date: Mon May 3 19:33:15 2010
> InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
> ProcEnviron:
> LANG=en_US.utf8
> SHELL=/bin/bash
> SourcePackage: apt
>
> To unsubscribe from this bug, go to:
> https:/
>
Victor Vargas (kamus) wrote : | #47 |
Please could somebody check if this behaviour is still occurring in the latest release included in Ubuntu Maverick Beta? here all works fine for me under Lucid/Maverick (behind a PFSense).
Changed in apt (Ubuntu): | |
importance: | Undecided → Medium |
status: | Confirmed → Incomplete |
YeagDeeDs (yeagdeeds) wrote : | #48 |
I have not noticed the error and I am running the latest build as of 09.13.2010.
Joan Teixidó (joanteixido) wrote : | #49 |
Hi,
i had the exactly same problem descripted in this post. I resolved it by commenting this line in source.list:
deb http://
I'd a lot of 3rd party sources but only when i enable this one (virtualbox..) the apt can't resolve any host (exactly like described in the post)...
Ubuntu 10.04
Joan
Drew Shetley (amshetle) wrote : | #50 |
This was an issue for me in Lucid, and now Maverick. It seems to only affect my 3rd party repos, but that could be a red herring.
Marcos Alano (mhalano) wrote : | #51 |
The problem is relatad with port number. Eg.:
ping linux.dropbox.
ping linux.dropbox.com (works fine)
The difference is ":http" part, used to indicate port number.
IMHO is a problem with resolve address code. Its resolving <address>:<port>, but the correct is only resolve <address>. Thanks.
Drew Shetley (amshetle) wrote : | #52 |
Is there an easy fix or workaround?
On Tue, Oct 12, 2010 at 12:20 PM, Marcos Henrique Esteves Barbosa <
<email address hidden>> wrote:
> The problem is relatad with port number. Eg.:
> ping linux.dropbox.
> ping linux.dropbox.com (works fine)
> The difference is ":http" part, used to indicate port number.
> IMHO is a problem with resolve address code. Its resolving
> <address>:<port>, but the correct is only resolve <address>. Thanks.
>
> --
> Apt has strange network errors when updating indexes.
> https:/
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “apt” package in Ubuntu: Incomplete
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random selection of
> indexes (from 0 to 3, usually). The errors look like this:
>
> W: Failed to fetch
> http://
> (-5 - No address associated with hostname)
>
> W: Failed to fetch http://
> (-5 - No address associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old ones
> used instead.
>
> It is usually about two that fail, though sometimes more, and sometimes
> none. I just have a wireless router attached to a cable modem, standard home
> broadband. If I run 'update' a couple times, it will get everything. Apt has
> no problem downloading or installing packages. I can run 'ping' against
> these servers with zero packet loss.
>
> If I do a single update with errors and then try to install packages,
> sometimes it warns me that packages are unsigned.
>
> Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't an
> upgrade problem, I backed up my personal files, and installed 10.04 from
> scratch. The problem persisted with the default Ubuntu 10.04 install.
>
> My laptop is a Compaq CQ60 215dx.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 10.04
> Package: apt 0.7.25.3ubuntu7
> ProcVersionSign
> Uname: Linux 2.6.32-21-generic x86_64
> Architecture: amd64
> Date: Mon May 3 19:33:15 2010
> InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release amd64 (20100429)
> ProcEnviron:
> LANG=en_US.utf8
> SHELL=/bin/bash
> SourcePackage: apt
>
> To unsubscribe from this bug, go to:
> https:/
>
Manfred Hampl (m-hampl) wrote : | #53 |
@Marcos Henrique Esteves Barbosa:
As far as I can see, that is not the problem.
When digging through the source I see that the syntax <address>:<port> is only used in the error message, but not in the real command that is executed to access the network.
By the way, I got that error myself a few weeks ago, but without doung anything a few days later the download and upgrade worked well without error (both on lucid).
MH
ubunturox (ubunturox-kk) wrote : | #54 |
Changing the DNS server worked like Thomas Rolland Butler mentioned
Austin Jones (ajones345) wrote : | #55 |
I have recently installed 10.10 Desktop on my netbook, which has had this exact problem before on the past 4 versions of Ubuntu. I am disappointed to see that after 4 entire versions this problem still has not been fixed. Ubuntu team, please focus on bugs, then do features.
maung.perfect (phopyaesone) wrote : | #56 |
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
W: Failed to fetch http://
jacobika (jacobtstewart) wrote : | #57 |
I was having this issue after switching from manually specifying my DNS sever in Network Manager to having it obtain the DNS address from my Belkin wireless router. I'm running Ubuntu 10.10. The fix for me was to switch back to manually specifying the DNS servers in Network Manager, as suggested in comment #18 above. If you're not sure of your DNS server, you can either look on your router's config page or use Google's public DNS servers (8.8.8.8 and 8.8.4.4).
tellapu (tellapu) wrote : | #58 |
The same issue with me, DNS Server change (Network Manager) to the OpenDNS 208.67.220.220 solved it for now.
mysticzzz (mysticzzz) wrote : | #59 |
Why is this set to medium importance? I am just retrying ubuntu on a new laptop, and installed it on a friends crashed laptop. This is a dealbreaker, As is the fact UBUNTU bug tracking and reporting leaves much to be desired. A simple search on google reveals 194 separate websites storing data on this exact error. Obviously this shows how poorly bug reporting procedures are related to the UBUNTU user. I have used numerous versions of UBUNTU and have never had a problem this severe. Inability to install software from the "Ubuntu software center" is not a Medium problem.
Kamil Kaczmarczyk (kamkacz) wrote : | #60 |
It seems apt simply has a too short timeout for dns queries. My router is apparently pretty slow at processing them so apt floods it easily. Passing direct dns addresses through DHCP or setting them directly on the computer solved the problem for me.
Gurubie (gurubie) wrote : | #61 |
Mysticzzz is right on target. I haven't had this problem with Ubuntu 10.04 (same router, same settings); now, upon installing the newer release, me and a bunch of folks are getting these apt errors.
So much, for the SEVERAL people, I just talked in to installing Maverick. I, at least thought we could depend on regular upgrades. This was a major reason that we could honestly say, it's better. Have we let the fox in the hen house, or what? I just don't know.
I'm having the same problem!
ricardisimo (ricardisimo) wrote : | #62 |
- screenshot - Motorola router page Edit (43.5 KiB, image/png)
Having the same problem, and I've tried any variation on the posted suggestions above, but quite honestly, when I go to my router's address at 192.168.7.254 (by the way fellow newbs, your router's IP address is to be found in the file /etc/resolv.conf) I get no useful options. I'm going to try posting the screenshot with the menu items for my Motorola router. If anyone can give some I dea as to where to go from here, please let me know.
My issue appears to be with the third-party repos as well. Although NOTHING is upgrading, at least the Ubuntu main repos appear to be updating. Any more permamnent fix for this in the works? Is it really a timeout issue? Can't that be lengthened?
maung.perfect (phopyaesone) wrote : | #63 |
I'm sick with, this bug. Somebody kill it.
Why was they put Importance to Medium?
I can't update my apt for a long time.
Dan (dorjedan) wrote : | #64 |
ricardismo / maung, have you guys tried changing your DNS via "network manager". This appears to solve the issue (did for me and others). I wasn't able to update it in my router but going the network manager route did the trick.
Sam Hiatt (shiatt) wrote : | #65 |
Still over 7 months after the OP, I too am experiencing the same problem.
Luckily the suggestion in post #36 seems to have SOLVED it, but this is ONLY A WORKAROUND. I also have AT&T (ugh!) and it seems AT&T's DNS servers are just too slow. I was able to manually set the Google DNS servers on my AT&T 2wire router, which is nice because this will solve the same problem for all machines on my network now. ... Hopefully this change will have the additional side effect of increasing my overall network latency (which previously sucked!).
But still... APT should definitely tweak it's settings to allow longer timeouts, or to retry a couple times after failure. No other network-connected software had problems with AT&T's crappy latency, so clearly APT is failing to accommodate such poor connections.
Does anyone out there know if these parameters are configurable? If so, how, and where? ... and if such a fix works, why aren't those defaults correctly configured in Maverick?
Drew Shetley (amshetle) wrote : | #66 |
Maybe I don't have the same problem after all, but all of my error messages
were due to PPAs. I deleted those PPAs, and redownloaded them, making sure
to get the GPG key for each, and that solved the problem for me. This bug
is incredibly annoying. Best of luck to everyone.
On Thu, Feb 17, 2011 at 1:08 PM, Sam Hiatt <email address hidden> wrote:
> Still over 7 months after the OP, I too am experiencing the same
> problem.
>
> Luckily the suggestion in post #36 seems to have SOLVED it, but this is
> ONLY A WORKAROUND. I also have AT&T (ugh!) and it seems AT&T's DNS
> servers are just too slow. I was able to manually set the Google DNS
> servers on my AT&T 2wire router, which is nice because this will solve
> the same problem for all machines on my network now. ... Hopefully this
> change will have the additional side effect of increasing my overall
> network latency (which previously sucked!).
>
> But still... APT should definitely tweak it's settings to allow longer
> timeouts, or to retry a couple times after failure. No other network-
> connected software had problems with AT&T's crappy latency, so clearly
> APT is failing to accommodate such poor connections.
>
> Does anyone out there know if these parameters are configurable? If so,
> how, and where? ... and if such a fix works, why aren't those defaults
> correctly configured in Maverick?
>
> --
> You received this bug notification because you are a direct subscriber
> of the bug.
> https:/
>
> Title:
> Apt has strange network errors when updating indexes.
>
> Status in “apt” package in Ubuntu:
> Incomplete
>
> Bug description:
> Binary package hint: apt
>
> If I update the apt indices either in Synaptic Package Manager, Update
> Manager or from the command line, it fails to download a random
> selection of indexes (from 0 to 3, usually). The errors look like
> this:
>
> W: Failed to fetch http://
> security/
> 'security.
>
> W: Failed to fetch
> http://
> happened resolving 'archive.
> associated with hostname)
>
> E: Some index files failed to download, they have been ignored, or old
> ones used instead.
>
> It is usually about two that fail, though sometimes more, and
> sometimes none. I just have a wireless router attached to a cable
> modem, standard home broadband. If I run 'update' a couple times, it
> will get everything. Apt has no problem downloading or installing
> packages. I can run 'ping' against these servers with zero packet
> loss.
>
> If I do a single update with errors and then try to install packages,
> sometimes it warns me that packages are unsigned.
>
> Originally, I had upgraded from Ubuntu 9.10, so to make sure it wasn't
> an upgrade problem, I backed up my personal files, and installed 10.04
> from scratch. The problem persisted with the default Ubuntu 10.04
> install.
>
> My laptop is a Compaq CQ60 215dx.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 1...
Sam Hiatt (shiatt) wrote : | #67 |
Yeah, if your 3rd party PPAs have outdated urls then you'd probably get a similar error message.
In my case I was getting the "No address associated with hostname" error on core packages, and it was occurring randomly, or random packages would trigger that error on each apt-get update attempt, then a different random set of packages would fail on the next try.
My symptoms definitely sound like APT is configured with a too short timeout period.
David Grooms (viadream) wrote : | #68 |
I have two different slices on slicehost.com (great host!), two different domains. Ubuntu 10.04 on each. I installed Ruby enterprise edition, mysql, and lots of good stuff on both using aptitude or apt-get. All was fine in paradise.
Then, I had to use a remote email system (at work) for one of the slices. This involved switching the dns to use the nameserver zones set up at work and they pointed back to the www .com pages on the slices. This was a work around, but it worked. Among other things to make the switch, I changed the resolv.conf to point to the new nameservers.
Several weeks latter, i first noticed, and for weeks, the apt-get failed only on the slice I had changed giving lots of errors when trying to use aptitude update/install or apt-get update/install, but the other slice was fine. Errors as described above:
Something wicked happened resolving 'archive.
I resolved this by adding the google public nameservers to the resolv.conf right after the two name servers from work and rebooting:
nameserver 8.8.8.8
nameserver 8.8.4.4
Hope this helps!
John Smith (anonymous-contributor) wrote : | #69 |
I have experienced this "wicked" behaviour when setting up firewall rules on a machine connecting to the Internet through a DSL modem. I believe that I have found a fix but I'm not completely certain.
The following conditions apply:
Clean install of Ubuntu 10.10
FreeNX installed per instructions on https:/
Guarddog 2.6.0 installed
The following seems to have fixed it for now:
Allow http(s) served from the internet zone to local zone (required by apt-get.)
Allow DNS served from the internet zone to the local zone (for remote DNS server.)
Allow DNS served from the intranet zone (192.168.1.0/24) to local zone (for the DSL modem's DNS server.)
Allow SSH served from the local zone to the intranet zone (for FreeNX.)
Enable DHCP on local Ethernet adapter.
Hope this helps too!
Tommy Trussell (tommy-trussell) wrote : | #70 |
This happened to me today, and I believe my router was misbehaving. I'm running a WRT54GL with a recent version of DD-WRT. I know from experience this WRT54GL unit gets weird after 1 to 2 weeks of uptime, so I have it set to reboot every night at 3am. HOWEVER for some reason it showed an uptime of 9 days, so I manually rebooted it. Now all seems fine.
It would be nice if the "something wicked" error were a bit more helpful, but on MY humble network, I think this means it's probably about time to buy a new router....
Julian Andres Klode (juliank) wrote : | #71 |
Not a bug. The error message is correct and provides all information required to find the cause of the issue in a straight-forward way.
Changed in apt (Ubuntu): | |
status: | Incomplete → Invalid |
Jason (jason-iphone-garbage) wrote : | #72 |
Sorry but this is not invalid. This is a bug.
For everyone who is having network issues, DNS, NAT, http proxy, etc. etc. then the error message is correct.
However there is something else going on here which affects users with correct network settings.
You can confirm this is not a network problem by using wget to directly access the URLs for which apt-get cannot download release info.
E.g.
wget http://
This will succeed, while "sudo apt-get update" complains about the same URL.
I have seen this in Ubuntu 10.04 and probably other releases. (Maybe 9.x and 10.10 ?)
This is somehow related to third party repositories.
The only workaround I have found is to disable all "Other Repositories"
To do this I go to
System -> Administration -> Synaptic Package Manager
Then go to
Settings -> Repositories
Go to the "Other Repositories" tab and uncheck everything.
Then select the "Close" button. This will give you a message about using the "Reload" button. You can then do that from within the Synaptic Package Manager UI.
Not sure if the "Reload" button is the UI equivalent, but you should now be able to run "sudo apt-get update" from the CLI
This workaround "fixes" the problem but I wonder how do I know if packages from my third party repos are up to date?
Changed in apt (Ubuntu): | |
status: | Invalid → New |
Julian Andres Klode (juliank) wrote : | #73 |
All comments in the bug indicate one of the following:
(a) bug in libc and fixed now
(b) temporary network issues
(c) misconfiguration
At no point did any comment indicate a bug in APT. Unless you can show that
sudo debootstrap natty natty; sudo chroot natty apt-get update
(and then bisect for maverick, lucid) does not work, please do not comment further
on this issue. Or try reproducing it with a chroot of your own release instead of natty
and see if it works, thus ruling out (c).
This bug turned into a network setup support channel, which is not the intention
of bugs. I do not want to hear anything about clicking anything, whether $RANDOM
programs work, or which networks are used. Unless someone serious can reproduce
the bug in a clean natty environment, there is no bug.
Changed in apt (Ubuntu): | |
status: | New → Incomplete |
Jason (jason-iphone-garbage) wrote : | #74 |
> All comments in the bug indicate one of the following:
> (a) bug in libc and fixed now
> (b) temporary network issues
> (c) misconfiguration
I'm not sure that my comment indicates one of those.
> At no point did any comment indicate a bug in APT. Unless you can show that
>
> sudo debootstrap natty natty; sudo chroot natty apt-get update
>
> (and then bisect for maverick, lucid) does not work, please do not comment further
> on this issue. Or try reproducing it with a chroot of your own release instead of natty
> and see if it works, thus ruling out (c).
This did not occur in natty (was it fixed there?)
This is also not an end user use case. I shouldn't have to chroot anything to recreate a bug.
> This bug turned into a network setup support channel, which is not the intention
> of bugs.
Oh I agree, and that is what is so frustrating trying to find a workaround or fix for this issue.
Almost every thread I find on this issue digresses into router configurations, DNS, NAT, etc etc which is not the issue here.
My network configuration is correct. The URLs apt complains about can be accessed as seen in my wget example above.
> I do not want to hear anything about clicking anything,
That's unfortunate.
I am trying to transition from an rpm based linux distribution to Ubuntu so I am unfamiliar with much of how apt works.
The workaround I posted above with UI steps for removing 3rd party repositories from apt was intended to help users affected with this same issue.
Based on your comments I will assume you are very familiar with the Ubuntu / synaptic / apt framework. You should then be able to understand what these use case steps through the UI do at a lower level to the apt configuration. And that might possibly help you to understand why removing 3rd party repositories through the UI steps I outline above would workaround the issues seen both through the GUI and through the apt CLI.
I posted here only hoping that it might help others experiencing the same issues with 3rd party repos and the misleading error messages indicating that apt cannot download from URLs (not from the 3rd party repos) which clearly can be resolved and accessed.
> whether $RANDOM
> programs work, or which networks are used.
> Unless someone serious can reproduce the bug in a clean natty environment, there is no bug.
I can appreciate that you feel I am not someone serious.
Still apt shouldn't choke because 3rd party repos make it unable to access the main Ubuntu repos, and subsequently complain with inaccurate error messages about the URLs it thinks it cannot access.
Julian Andres Klode (juliank) wrote : | #75 |
My goal is to check whether it works for the affected people in a minimal environment. If it does, it is a misconfigured system. If it doesn't, then it is either network or bug. If it is a bug, we need to find out when it broke and when it was fixed.
As soon as graphical frontends enter the picture, things become more complex.
Once you have seen that a minimal environment works, change it to become more like your environment until the problem arises. Then you know the cause of the problem.
Sam Hiatt (shiatt) wrote : | #76 |
> Still apt shouldn't choke because 3rd party repos make it unable to access
the main Ubuntu repos, and subsequently complain with inaccurate error
messages about the URLs it thinks it cannot access.
I agree.
My experience with this issue suggests that my crappy internet connection is
causing apt to time out when it can't reach the repos quickly enough. But
even if my connection is to blame, apt should be more resilient. Maybe by
simply retrying a couple of times.
With regard to GUIs, I experience this issue in the CLI as well as in
Synaptic.
David Brown (real-mahjong) wrote : | #77 |
I'm having this problem with 12.04, which has been running and updating quite happily for several months. I Eventually applied the workaround described in http://
David Brown (real-mahjong) wrote : | #78 |
Addendum to last comment:
For clarity, I should have said that the solution suggested in the link was to set the DNS to 208.67.222.222, 208.67.220.220 directly.
Kamil Kaczmarczyk (kamkacz) wrote : | #79 |
@up
Yeah, that's basically the solution which has been suggested here for, like, 2 years? The problem with the fact this solution works it is that it only proves that the issue is external, that it's your router's built-in DNS server's fault - it's flooded with so many requests coming all at once. Apt might possibly implement a fallback mechanism of some sort, like detecting such a situation and polling the servers in sequence, but, well, I guess it's of low priority.
david6 (andrew-dowden) wrote : | #80 |
Same issued just occured under 13.04 (raring), for Australia.
This appears to be a DNS collision issue ..
david6 (andrew-dowden) wrote : | #81 |
Disregard.
Caused by medibuntu.org (no DNS record), due to no longer being maintained.
flickerfly (josiah-ritchie) wrote : | #82 |
I have been experimenting with the transparent HTTP proxy on my Watchguard XTM firewalls. The HTTP proxy was catching the update requests over HTTP because they had no content_type. I added the following domain exceptions to resolve this: archive.
I could also have simply turned off checking for the lack of content type, but I'm trying to learn how to work with this thing. (Thanks for the practice.)
dino99 (9d9) wrote : | #83 |
Support for this version has ended
Changed in apt (Ubuntu): | |
status: | Incomplete → Invalid |
I confirm this problem as well. I just did a freash 10.04 install (64bit). us.archive. ubuntu. com lucid-updates/ universe Packages us.archive. ubuntu. com lucid-updates/ universe Sources us.archive. ubuntu. com lucid-updates/ multiverse Packages us.archive. ubuntu. com lucid-updates/ multiverse Sources us.archive. ubuntu. com/ubuntu/ dists/lucid/ Release. gpg Something wicked happened resolving 'us.archive. ubuntu. com:http' (-5 - No address associated with hostname)
....
Hit http://
Hit http://
Hit http://
Hit http://
Fetched 3,638B in 11s (315B/s)
W: Failed to fetch http://