empty local package list

unicorn

I run pkgin upgrade yesterday, system froze to death. When I run it again today pkgin said empty local package list. How to resume upgrade?

unicorn what's the url in repositories.conf ?

rvp

unicorn I run pkgin upgrade yesterday, system froze to death.

Did you run fsck on your filesystems after this happened?

unicorn

URL

unicorn Is this your own mirror? Are there updates available? pkg_summary says 2023-03-28 15:31

unicorn

It's not my mirror, it's a public mirror. I changed to this mirror last night because the download speed is very slow.
I get a fast download speed with this mirror.

unicorn did you run pkgin update after you changed to this repository and before you try to upgrade?

unicorn

I can't remember. I had downloaded all packages, but the system was dead when upgrading packages, maybe because I touched the cable caused the system dead.

unicorn

unicorn pkgin upgrade all give the same error

Could you please be more specific? What error do you get?

unicorn maybe because I touched the cable caused the system dead.

Hmm ... I wonder if the package database got corrupted? Check /usr/pkg/pkgdb

unicorn

I tried different mirrors, pkgin update works, pkgin upgrade all give the same error. How to solve this problem?

unicorn

the title is the error message. /usr/pkg/pkgdb have all the packages. When switch to cdn.netbsf.org, pkgin update can't find pkg_summary.gz

@unicorn
If pkgdb is indeed messed-up, run pkg_admin rebuild.
See, man pkg_admin

unicorn

pin
pkg_admin rebuild can't be executed, aborted.

unicorn

rvp
can't run fsck on /, as file system is mounted read-write on /

rvp

unicorn can't run fsck on /, as file system is mounted read-write on /

Boot into single-user mode, use /rescue/sh as the shell, then run:

fsck -f /

This is standard stuff, mate. Every Unix user should know this.

unicorn the title is the error message. /usr/pkg/pkgdb have all the packages.

You're not explaining this very well, and we can't see what's happening on your system.

Can you upload these 2 files somewhere?

/var/db/pkgin/sql.log
/var/db/pkgin/pkg_install-err.log

(The first one, sql.log, may not exist.)

What're the outputs of these commands:

ls -l /usr/pkg/pkgdb/*.db /var/db/pkgin/
cat /usr/pkg/etc/pkgin/repositories.conf
pkgin list > /tmp/pkgin.log
pkg_info   > /tmp/pkg_info.log

Upload the 2 /tmp/*.log files also.

unicorn

rvp

I had those four log files stored in the system, but I can't open firefox, and I can't transfer them to my phone as I can't open ftp application. I can't transfer files using mtp either.I had run fsck, there's no error found.

unicorn Does git still work? If so create a repo on another machine, clone it on the problem system and commit the files.

An alternative is katbin-cli but, if you don't already have it installed, there's no help.

rvp

unicorn I can't transfer them to my phone as I can't open ftp application. I can't transfer files using mtp either.

This is not difficult. You can:

Zip all the files up and copy them to a DOS-formatted USB drive, then use another system.
Zip everything up (incl. the output of the commands), then use nc to transfer files. You can get nc even on Windoze, I think.
Maybe use Emacs? It already has everything else--incl. a browser, so why not a file xfer service too?

Jay

unicorn if you have running internet connection try this http://termbin.com/

rvp

unicorn I had run fsck, there's no error found.

OK, but, if /usr or /var are separate filesystems, then you'll check them also.

Well, I can see errors in your logs, but, they're mostly from last year:

sql.log

[...]
SQL error at 2022-05-05 02:52:44 CST
   errmsg: statement aborts at 13: [INSERT INTO LOCAL_PROVIDES (PKG_ID, LOCAL_PROVIDES_PKGNAME) VALUES (116,'/usr/pkg/lib/gstreamer-1.0/libgstsubparse.so');] attempt to write a readonly database
    query: INSERT INTO LOCAL_PROVIDES (PKG_ID, LOCAL_PROVIDES_PKGNAME) VALUES (116,'/usr/pkg/lib/gstreamer-1.0/libgstsubparse.so');
SQL error at 2022-05-05 02:52:44 CST
   errmsg: statement aborts at 13: [INSERT INTO LOCAL_PROVIDES (P

pkg_install.log

[...]
---May 04 21:32:25: refreshing gdk-pixbuf2-2.42.6nb1...
g_module_open() failed for /usr/pkg/lib/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-svg.so: /usr/pkg/lib/libpangoft2-1.0.so.0: Undefined PLT symbol "hb_font_set_synthetic_slant" (symnum = 7)
[1]   Segmentation fault (core dumped) ${GDK_PIXBUF_QUERY_LOADERS} >${GDK_PIXBUF_LOAD...
[...]
---May 05 02:41:56: refreshing serf-1.3.9nb2...
pkg_add: Read error for lib/libserf-1.so.1.3.0: Lzma library error: Corrupted input data
pkg_add: Updating of serf-1.3.9nb2 to serf-1.3.9nb2 failed.
pkg_add: Remember to run pkg_admin rebuild-tree after fixing this.
pkg_add: Couldn't remove /usr/pkg/lib/pkgconfig/serf-1.pc: No such file or directory
pkg_add: 1 package addition failed
[...]
---May 28 18:48:40: refreshing gdk-pixbuf2-2.42.6nb1...
pkg_delete: Couldn't remove package directory in `/usr/pkg/pkgdb/gdk-pixbuf2-2.42.6nb1'
pkg_add: command failed: /usr/pkg/sbin/pkg_delete -K /usr/pkg/pkgdb -p /usr/pkg  gdk-pixbuf2-2.42.6nb1
pkg_add: 1 package addition failed
---May 28 18:48:41: refreshing gdbm-1.23...
pkg_add: /usr/pkg/pkgdb/gdk-pixbuf2-2.42.6nb1/+CONTENTS: No such file or directory
---May 28 18:48:41: refreshing gd-2.3.3nb1...
pkg_add: /usr/pkg/pkgdb/gdk-pixbuf2-2.42.6nb1/+CONTENTS: No such file or directory
---May 28 18:48:41: refreshing gcc7-7.5.0nb5...
[...]

Many other there...

More relevent is the output of pkg_info which shows definite corruption:

[...]
xfce4-power-manager-4.16.0nb2 Xfce power manager
py310-expat-3.10.9nb1 /* XPM */
static char * to

Right. The easiest thing to do is reinstall after wiping everything:

rm -rf /var/db/pkgin /usr/pkg
export PKG_PATH='http://cdn.NetBSD.org/pub/pkgsrc/packages/NetBSD/amd64/10.0/All'
pkg_add pkgin
echo 'http://cdn.NetBSD.org/pub/pkgsrc/packages/NetBSD/$arch/10.0/All' > /usr/pkg/etc/pkgin/repositories.conf
pkgin update
pkgin install [...]

Replace 10.0 everywhere with 9.3 or 9.2 or whatever version of NetBSD you have.

Otherwise, you can try to salvage your current install:

Move old DB out of the way. (You can remove these later...)

mv /var/db/pkgin /var/db/pkgin.orig
mv /usr/pkg/pkgdb/pkgdb.byfile.db /usr/pkg/pkgdb/pkgdb.byfile.db.orig

Fix corrupted files.
Running pkg_info causes it to read the +CONTENTS file in each package directory in /usr/pkg/pkgdb/. You can see this is failing because (at least one of) the files is corrupt:
```
py310-expat-3.10.9nb1 /* XPM */
static char * to
```
I suspect that some of the files in /usr/pkg/pkgdb/py310-expat-3.10.9nb1/ are bad. So, download that tarball from here (use correct version please), then extract only the +-files, over-writing the possibly corrupt files in the /usr/pkg/pkgdb/<package name>/ directory.
Run, pkg_info again, and check for other corrupt packages. Repeat as necessary.

Rebuild database
When the pkg_info output looks OK, try rebuilding the database:

pkg_admin rebuild

If there are any errors here, then note which packages are corrupted, do Step 2 for each of them, then rebuild again until the rebuild succeeds without errors.

Check packages
By this point the database should be consistent. Now check all the installed package files:

pkg_admin check

If any files fail in their checksums, remove the package and reinstall:

pkgin clean
pkgin update
pkgin rm [...]
pkgin autoremove
pkgin in [...]

This should make things reasonably consistent. @pin can chime in with corrections...

unicorn

rvp

I can use emacs, how to do?

unicorn

I take some pictures
pic

unicorn please use one of the methods suggested by @rvp, @Jay or, myself.

Don't know about the others but, I'm not looking at logfiles as pictures, sorry.