"2321" packages to upgrade . . . on reboot another "204" yet to go?? Why???

English Open Chat
1

Folks:

I’ve posted on the forum previously about the large number of package upgrades that will show up . . . in what is supposedly a rolling distro, and yet the behavior continues . . . . This morning, a mere week after the last upgrade was done in TW, the updater showed “2321” packages available to upgrade . . . should be “major system upgrade” . . . essentially moving from one kernel to the next iteration, 5.17.4 to 5.17.5???

I blithely began the process in the GUI console at 8:01 AM my local time . . . I went to eat breakfast in the other room. When I came back 20 mins later the computer had shut itself down??? I restarted the machine and logged into a TTY . . . seems like the packages up to “1803” were “in cache.” And the retrieval process went on from there . . . . The cursor returned, “Please reboot for best outcomes,” so I did–finish time on the upgrade was 10:01 AM. TW is the grub controller for my numerous linux installs and on reboot it took another “5 minutes” to get passed the “purging old kernels” command and into the GUI.

After I logged into the GUI and launched my browser . . . a few minutes later the package updater shows “204 packages available to upgrade”??? So, we have two hours of computer time to download and install 2321 packages . . . and minutes later yet another 204 yet to go???

My question, the same one I asked the last time TW had such large package upgrades, is why? Why, in a rolling distro do these balloon package upgrades show up, taking substantial computer time to run through, after which . . . the GUI is “the same as it ever was” . . . superficially it is identical in spite of all of that upgrading.

In very stark contrast, I recently upgraded my Lubuntu install from “jammy” to “kinetic” . . . and there apt found “192 packages available to upgrade.” I wonder aloud whether the whole system even has “2321 packages,” let alone needing to download and upgrade all of them to make a system change from, e.g., 22.04 to 22.10. I also have a Manjaro, rolling edition . . . usually package numbers are low . . . rarely getting into double digits–that to me is what “rolling” means.

I’m not a new user of TW, historically there wasn’t this huge number of packages to download and install, it’s only been in the last couple of years that every few weeks or months these huge numbers show up in the tubes, no apparent way to break them into “biteable” chunks??? Is this really necessary, or is it just “easier” for the upstream folke? Please advise.

2

There were forwarnings. E.g. https://forums.opensuse.org/showthread.php/569897-GCC-12-Becoming-Default-Compiler-in-Tumbleweed

And this one: https://forums.opensuse.org/showthread.php/569872-Incoming-Full-rebuild-of-Tumbleweed-(GCC12)

And many people have already swallowed it.

3

Yep I agree, it’s a total BS !

I ran the update last saturday (2750 packages). Took several hours ! (I run 32 bit tw on an old netbook with external monitor), only to discover it would not boot into gui (xfce) after the update. What a piece of xxxx !

The x11 drivers were updated, but not loading acc. to X11 log ! How is that possible ?

I have reverted to a pre update snapshot and updated xfce through yast. The system seems stable again now.

I will hold off updating the other 2500 packages or so for now until I feel more confident.

This updating of 2500 packages or more every week is not very user friendly to say the least it’s a plain PIA !

4

Hi
Fiirst off, I suggest you tone it down, it won’t be tolerated in this Forum!

Follow the mailing list, users were forewarned. It IS to be expected glibc updates, gcc updates cause rebuilds and big updates.

5

FYI,

I have tumbleweed in a vm that had 3300 updates today since last updated last week. But, I also have arch linux, and rocky linux as vms. These do not have nearly as comprehensive a set of software installed as my tw vm. They each had over 1000 updates today since updating last week. I think it just happened that a bunch of updates were released concurrently by the developers.

I believe, If you were on a system other than openSUSE TW, you likely would have had all of the same updates.

tom kosvic

6

Don’t believe, check. TW has moved to a newer GCC, and everything will be rebuilt. Hence the huge update.

7

It doesn’t have to be bad. Today I upgraded a TW 20220218 to 20220515. There were 1177 transactions in just over 10 minutes to complete zypper dup. The PC is over 4 years old, but does run on NVME. It’s quick to boot too:

# systemd-analyze
Startup finished in 12.938s (firmware) + 30.248s (loader) + 1.234s (kernel) + 1.405s (initrd) + 2.830s (userspace) = 48.657s
graphical.target reached after 2.824s in userspace
8

It’s the nature of the beast
It’s nothing to be complaining about
If your network is limited, TW may not be the best choice for you

9

Distribution upgrades are not an issue. Host 6700k runs service dup:

[FONT=monospace]**6700K:~ #** systemctl cat dup.service dup.timer                                                     
**# /etc/systemd/system/dup.service**
[Unit] 
Description=Dist Upgrade 

[Service] 
ExecStartPre=/usr/bin/sleep 11 
ExecStart=/usr/bin/zypper dup --no-confirm 

**# /etc/systemd/system/dup.timer**
[Unit] 
Description=Systemd timer to update the system daily with zypper dist-upgrade  

[Timer] 
OnCalendar=daily 
Persistent=true 

[Install] 
WantedBy=timers.target 
**6700K:~ #**[/FONT]

Modest CPU time consumption:

[FONT=monospace]**6700K:~ #** journalctl -u dup  -q -g Consumed    
May 14 08:29:01 6700K systemd[1]: dup.service: Consumed 1.686s CPU time. 
May 15 03:07:34 6700K systemd[1]: dup.service: Consumed 2min 2.284s CPU time. 
May 15 06:40:24 6700K systemd[1]: dup.service: Consumed 40.367s CPU time. 
May 15 07:25:04 6700K systemd[1]: dup.service: Consumed 13.075s CPU time. 
May 16 05:49:57 6700K systemd[1]: dup.service: Consumed 15.634s CPU time. 
May 16 21:40:59 6700K systemd[1]: dup.service: Consumed 4min 39.034s CPU time. 
May 17 03:32:13 6700K systemd[1]: dup.service: Consumed 4.684s CPU time. 
**6700K:~ #**[/FONT]

[FONT=monospace][FONT=monospace]
[/FONT][/FONT]
Typical load times are: 711ms (kernel) + 2.058s (initrd) + 1.816s (userspace).

**6700K:~ #** journalctl -q -g 'Reached target Graphical Interface' -o short-monotonic  
    4.862609] 6700K systemd[1]: Reached target Graphical Interface. 
    4.470168] 6700K systemd[1]: Reached target Graphical Interface. 
    4.533455] 6700K systemd[1]: Reached target Graphical Interface. 
    5.048718] 6700K systemd[1]: Reached target Graphical Interface. 
    4.903398] 6700K systemd[1]: Reached target Graphical Interface. 
    4.586601] 6700K systemd[1]: Reached target Graphical Interface. 
   34.547371] 6700K systemd[1]: Reached target Graphical Interface. 
**6700K:~ #**

Occasionally purge-kernels.service delays start of X.

See also: https://forums.opensuse.org/showthread.php/566863-Tumbleweed-20220219-quot-Mega-Update-quot?p=3127226#post3127226

10

Not on Leap. I assume this will reach Leap users on upgrade to next version (15.4, or the one after that?). And they expect to get everything new on such a version upgrade anyhow.

TW is rolling! And when you choose for that, you should roll with it.

11

If a distribution’s GNU Compiler Collection (GCC) is updated to a newer major release – meaning that, the Languages being compiled move from a given major version to a newer major version then –

  • Due to the changes and additions made to the syntax of the Languages being compiled, EVERYTHING
    will be re-compiled, as a matter of course due to past (nasty) experiences experienced in the past by the community of people who program and maintain computer systems …

Hence, the (large) number of changed packages which have to be downloaded.

12

If the GCC release is changed from a given Leap release to a newer Leap release – regardless of a change to a newer point-release or a newer major-release (major-releases often suffer a GCC release update) – then, all the packages will be upgraded – even if the developer versions of the packages do not change.

  • If the GCC release doesn’t change then, only those packages which have newer developer versions will be upgraded …
13

I understand why, but the frequency of the recompiles is very often imo and especially on older hardware (32 bit) and/or lower bitrate connections the massive updates are a pain !

I’m holding off a 2500 package update atm after updating to latest xfce and kernel and everything is running just fine…:slight_smile:

If there is another recompile in 2 or 3 weeks you can do it again…

Cheers!

14

Although I sometimes get the feeling that people who make Linux and Windows, think I install the os just to update it, especially on Tumbleweed you don’t have to update.

Nobody is forcing you.

I run windows 10 in a virtual machine, and if I could, I would turn off updates all together there.
Typical windows though, that is not so easy. Me and microsoft, don’t seem to agree on who owns this pc.

On tumbleweed how ever, I only have to click one red cross, if I am not in the mood for updates.
Yes it sometimes is annoying, but I was warned upfront about this. (Tumbleweed is for developers, who need the latest of the latest)

Because I am also a gamer, and how much the graphic drivers have improved (thank you Valve, for bringing steam to Linux) in the last months, I also benefit from having the latest of latest kernel and video drivers.

If you wait long enough, you will be probably skipping updates, but I am 99% sure this won’t be an issue.

If you do not have a relative fast pc though, I can understand this is more then a little annoying.
Still every time, updating windows takes much longer then updating Tumbleweed over here.

What I noticed though, updating windows 10 in my virtual machine, takes more time, then updating the actual win10 install also on this pc.
Still much slower then updating Tumbleweed though.

My advise would be don’t upgrade. If your pc is doing all you want it to do, no need for updates. (especially in a VM) :slight_smile:

15

Hi
It’s your system, you can do what you want :wink: However it can lead to issues going forward, especially from a security perspective. One issue that users faced not too long ago was the rpm update, all of a sudden couldn’t update. Tumbleweed is meant to roll, if you can’t handle the process, then Leap is a better choice (Doesn’t help those 32bit systems still around though).

The most important thing IMHO is to keep an eye on the Factory Mailing list, especially after big updates to check on any issues, then proceed. You can always set to download only, get a few cached, then repeat, once happy do a zypper dup…

16

Folks:

I appreciate the comments . . . I had my options set for “reply instantly via email” and didn’t get any notifications on any of the posts on the thread . . . something further to complain about . . . . : - )

So, this morning is Debian Sid day, and running apt in Sid can be “interesting,” but not as interesting as TW . . . today Sid brought a very heady “350 packages” to upgrade . . . . Yesterday, in my Sys76 linux laptop where I have Gecko Plasma installed along with Pop! & Bookworm, Gecko showed “1657 packages” to upgrade, rather than the “2321” that I ran in TW . . . a substantial difference, for essentially the “same” system . . . that machine is substantially quicker than my '12 i7 desktop, but still, “less packages” . . . to process for the upgrade, etc.

It’s not that I have time to search the Blog to get a head’s up on “tsunami of packages headed your way” . . . I can see that when I run zypper . . . . The question is, “Why is it that other rolling systems I have installed are not having this huge number of upgrades every couple of weeks/months?”

Is their “rolling” substantially “under-funded” and “incomplete”?? Whether or not the under hood “engine” is completely “refreshed” with the 2321 packages, followed by another “204” to get it done . . . an OS GUI is an OS GUI . . . .

I like having the “freshest horsies” available, that’s why I have most of my systems installed are “rolling,” or “beta” or “Sid” . . . but TW is in the field of essentially “one” with these astronomical package upgrades . . . .

Linux is “like a shark . . . it has to move forward or die . . .” . . . but, TW is what could be termed, “high maintenance” . . . a very high maintenance system.

17

I think that @dcurtisfra’s post #12 above explains it well enough. What other distributions do with the phenomenon, I have no idea. Ask there. IMHO you should either you accept the TW concept or use something else. There is a third option, join the TW developers team to discuss and have influence on how they manage things.

18

May I correct you?

19

May I correct you?

  • Computing is like a shark.

Some inherent aspects of computing –

  • Writing computer code is the most difficult task that human beings have burdened themselves with.
  • Computer code is software – it can be changed but, mostly the changes are neither simple nor easy …
  • The languages used for coding are continuously evolving – they’re changing – continuously …
  • Small amounts of code only occur on small embedded systems – all other computing systems have large amounts of code.
  • Computing is not quite like a shark – it is inherently moving due to human endeavours and inspirations and thoughts – and therefore, it’ll only die (on this planet) when the last human being dies …
  • Computing possibly galactic – it’s conceivable that, other living entities somewhere in the Galaxy are also using computers …

[HR][/HR]Bottom line –

  • If you’re upset by the amount of resources consumed by computing, stop using computers …
20

@malcolmlewis:

OK, this is a helpful suggestion . . . . In my case, I’m like a shark . . . each day I’m swimming in a different OS, to avoid “death due to boredom” . . . so I’d still have to download/cache all of the packages in the single session, right?? Once I reboot out of TW the cache is cleansed . . . ???

@hcvv:

Less helpful suggestions here . . . I am an end-user rather than a dev. I am on the Lubuntu dev list-serve, but it is rare that anything gets posted there. In the case of TW, when I , as an end-user find some issue, I post about it–hence the post.

But, you are suggesting that I should ask, e.g., the Manjaro forum, “Hey guys, how comes Manjaro (rolling) has such precise and efficient package upgrades compared to . . . TW??? I mean can you explain why your system has substantially less packages to upgrade through a year?” Based upon my experience with that forum, similar to the Debian forums, I would expect nothing but abuse coming back . . . for asking such a question.

My question is posted here, where this “galactic sized package upgrades is happening . . . .” The other star systems in the galaxy are using less of my computer’s time and energy to maintain . . . .

Obviously, the conversation will now repeat, etc. It appears that I joined this forum in '16 . . . at that time “rolling” in TW had the similar experience of the other rolling editions, i.e., regular package upgrades in double digit or triple digit numbers. Only in the last couple of years have these huuuuggggeee package numbers hit the tubes. Was there no “GCC” upgrades back then???

My “introduction” to the linux user mentality was via a student back in the early/mid '00’s who showed me a PC that he installed linux on . . . he told me that he had another computer that “runs off of a 126 MB floppy, the whole system . . . 126 MB” he repeated for emphasis. At that time OSX Tiger was something like 4 GB . . . OSX expanded way beyond that lowly number . . . .