| View previous topic :: View next topic |
| Author |
Message |
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
 Posted: Wed Mar 27, 2024 1:31 pm Post subject: |
 |
|
| pietinger wrote: | I never recommend "--keep-going" ... because if there is an error it is better to find the reason for it and then correct the error before starting emerge again. If someone is afraid to start "emerge -e" again, then I want point to this brilliant solution from @Massimo B. =>
https://forums.gentoo.org/viewtopic-p-8821469.html#8821469 |
Yea, I ended up starting my world update without using --keep-going. That solution in the other thread is interesting...unfortunately neither of the machines I'm updating have eix installed.
In the event of, for example, a failure that I'm able to resolve, would re-running the same emerge command with --resume continue starting with the failed package?...and should the --resume be specified without also using --emtytree?
Thanks in advance!
Tom |
|
| Back to top |
|
 |
figueroa
Advocate
Joined: 14 Aug 2005
Posts: 3005
Location: Edge of marsh USA
|
| Posted: Wed Mar 27, 2024 2:13 pm Post subject: |
|
|
In the event of an --emptytree failure, I note the failure, then restart with "emerge --skipfirst --resume" where --skpfirst jumps over the failed package, and --resume restarts the incomplete emerge where it left off. I've heard that syntax is important such that --skipfirst should come ahead of --resume in order to have the right effect. If I'm wrong about that, it's a harmless wrong.
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi |
|
| Back to top |
|
|
figueroa
Advocate
Joined: 14 Aug 2005
Posts: 3005
Location: Edge of marsh USA
|
| Posted: Wed Mar 27, 2024 2:17 pm Post subject: |
|
|
| grknight wrote: | | figueroa wrote: | | Code: | | emerge --ask --emptytree --exclude "$(qlist -IC acct-) dev-build/libtool $(qlist -IC media-fonts/) sys-apps/baselayout sys-devel/binutils sys-devel/gcc sys-firmware/intel-microcode sys-kernel/gentoo-sources sys-kernel/linux-firmware sys-kernel/linux-headers sys-libs/glibc $(qlist -IC -- -bin) $(qlist -IC virtual/)" @world |
|
FWIW, you can replace most of those qlist commands with, e.g., virtual/* and the like in the exclude list. |
Thank you, grknight. That's good to know. I wasn't sure about using wildcards. I checked each of those qlists elements and they run fast and work as intended.
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi |
|
| Back to top |
|
|
grknight
Retired Dev
Joined: 20 Feb 2015
Posts: 1912
|
| Posted: Wed Mar 27, 2024 2:21 pm Post subject: |
|
|
| figueroa wrote: | | grknight wrote: | | figueroa wrote: | | Code: | | emerge --ask --emptytree --exclude "$(qlist -IC acct-) dev-build/libtool $(qlist -IC media-fonts/) sys-apps/baselayout sys-devel/binutils sys-devel/gcc sys-firmware/intel-microcode sys-kernel/gentoo-sources sys-kernel/linux-firmware sys-kernel/linux-headers sys-libs/glibc $(qlist -IC -- -bin) $(qlist -IC virtual/)" @world |
|
FWIW, you can replace most of those qlist commands with, e.g., virtual/* and the like in the exclude list. |
Thank you, grknight. That's good to know. I wasn't sure about using wildcards. I checked each of those qlists elements and they run fast and work as intended. |
Like everything with portage, this only works with a full category and * as the package. Partial names will not work |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Wed Mar 27, 2024 2:42 pm Post subject: |
|
|
| figueroa wrote: | | In the event of an --emptytree failure, I note the failure, then restart with "emerge --skipfirst --resume" where --skpfirst jumps over the failed package, and --resume restarts the incomplete emerge where it left off. I've heard that syntax is important such that --skipfirst should come ahead of --resume in order to have the right effect. If I'm wrong about that, it's a harmless wrong. |
Just to be clear, does the "emerge --skipfirst --resume" require any other options at all or the @world etc...or does that command as-is just resume based on the current resume list?
Thanks!
Tom |
|
| Back to top |
|
|
figueroa
Advocate
Joined: 14 Aug 2005
Posts: 3005
Location: Edge of marsh USA
|
| Posted: Wed Mar 27, 2024 3:46 pm Post subject: |
|
|
| tld wrote: | | figueroa wrote: | | In the event of an --emptytree failure, I note the failure, then restart with "emerge --skipfirst --resume" where --skpfirst jumps over the failed package, and --resume restarts the incomplete emerge where it left off. I've heard that syntax is important such that --skipfirst should come ahead of --resume in order to have the right effect. If I'm wrong about that, it's a harmless wrong. |
Just to be clear, does the "emerge --skipfirst --resume" require any other options at all or the @world etc...or does that command as-is just resume based on the current resume list?
Thanks!
Tom |
No, just "emerge --skipfirst --resume" and adding anything else may muck it up.
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Wed Mar 27, 2024 8:24 pm Post subject: |
|
|
| figueroa wrote: | | No, just "emerge --skipfirst --resume" and adding anything else may muck it up. |
That's what I thought. Thanks for confirming!
Tom |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Thu Mar 28, 2024 12:43 pm Post subject: |
|
|
| tld wrote: | | figueroa wrote: | | No, just "emerge --skipfirst --resume" and adding anything else may muck it up. |
That's what I thought. Thanks for confirming!
Tom |
Good thing I was prepared for this. I had the compile of MythTV fail because it needed a patch for binutils >= 2.41. I was able to add a patch to my existing user patches,
So in this case I just used "emerge --resume" without any skipfirst. What confuses me however is that, while the number of packages it resumed did exactly match the number I had remaining including mythtv, it doesn't appear it's in the same order that it was, because mythtv wasn't first. I'm still waiting for it to get to that.
Is there some command to show the current list of packages that emerge is using in order?? Thanks.
EDIT: So out of the 32 remaining packages to resume, mythtv came up 5th. I'd still be curious as to how to access the package list of the current emerge. There have been other times I wanted to determine that. Thanks!
Tom |
|
| Back to top |
|
|
logrusx
Advocate
Joined: 22 Feb 2018
Posts: 2407
|
| Posted: Thu Mar 28, 2024 3:30 pm Post subject: |
|
|
Emerge does not blindly follow the order of the packages remaining, it doesn't know if you did something in the mean time so it must calculate the dependency graph of the remaining packages to make sure nothing breaks and this group now has a different graph which may result in different order.
Best Regards,
Georgi |
|
| Back to top |
|
|
tld
Veteran
Joined: 09 Dec 2003
Posts: 1845
|
| Posted: Thu Mar 28, 2024 5:16 pm Post subject: |
|
|
| logrusx wrote: | Emerge does not blindly follow the order of the packages remaining, it doesn't know if you did something in the mean time so it must calculate the dependency graph of the remaining packages to make sure nothing breaks and this group now has a different graph which may result in different order.
Best Regards,
Georgi |
I see. I'd have to assume however, that if I were to have used --skipfirst it would have properly skipped the package that failed and NOT the one that ended up first in that new order(?).
Thanks!
Tom |
|
| Back to top |
|
|
figueroa
Advocate
Joined: 14 Aug 2005
Posts: 3005
Location: Edge of marsh USA
|
| Posted: Thu Mar 28, 2024 5:50 pm Post subject: |
|
|
| tld wrote: | | logrusx wrote: | Emerge does not blindly follow the order of the packages remaining, it doesn't know if you did something in the mean time so it must calculate the dependency graph of the remaining packages to make sure nothing breaks and this group now has a different graph which may result in different order.
Best Regards,
Georgi |
I see. I'd have to assume however, that if I were to have used --skipfirst it would have properly skipped the package that failed and NOT the one that ended up first in that new order(?).
Thanks!
Tom |
Yes it does. Portage is very smart.
_________________
Andy Figueroa
hp pavilion hpe h8-1260t/2AB5; spinning rust x3
i7-2600 @ 3.40GHz; 16 gb; Radeon HD 7570
amd64/23.0/split-usr/desktop (stable), OpenRC, -systemd -pulseaudio -uefi |
|
| Back to top |
|
|
saturnalia0
Apprentice
Joined: 13 Oct 2016
Posts: 158
|
| Posted: Thu Mar 28, 2024 6:51 pm Post subject: |
|
|
Thanks for the tip, I upgraded today and it went flawlessly. Took under 6h on my i5-7400 (4 cores at 3.0GHz), so less than a day's work 
This was the command I used:
| Code: |
emerge --ask --emptytree --exclude "sys-devel/gcc sys-devel/binutils sys-libs/glibc dev-build/libtool linux-firmware intel-microcode gentoo-sources linux-headers baselayout" @world
|
790 packages rebuilt. |
|
| Back to top |
|
|
|
Portage & Programming
