INFO:
You're in good shape at 5.8 GB free. The tar file can go straight from where it is. dbv6admin can already read /var/www/html/rbb-rhel8-repo-20261117.tar.gz, so it doesn't need copying to /tmp on B first. Only extra-rpms (160 MB) has to go through /tmp, because dbv6admin can't get into /root.
On B, as root
1. Put extra-rpms where dbv6admin can reach it, and take the tar file's fingerprint
cp -r /root/releases/2026.11/extra-rpms /tmp/
chown -R dbv6admin:dbv6admin /tmp/extra-rpms
cp /root/os_patch_wizard_1.5.2.sh /tmp/
chown dbv6admin:dbv6admin /tmp/os_patch_wizard_1.5.2.sh
sha256sum /var/www/html/rbb-rhel8-repo-20261117.tar.gz
exit
Write down the code sha256sum prints. You'll compare it on A. Skip the two wizard lines if A already has 1.5.2. To check, run grep SCRIPT_VERSION= /root/os_patch_wizard*.sh on A.
On B, as dbv6admin
2. Send everything to A's /tmp
scp /var/www/html/rbb-rhel8-repo-20261117.tar.gz dbv6admin@<A-master>:/tmp/
scp -r /tmp/extra-rpms dbv6admin@<A-master>:/tmp/
scp /tmp/os_patch_wizard_1.5.2.sh dbv6admin@<A-master>:/tmp/
On A
3. Log in as dbv6admin, check the space, then become root
df -h /
su -
You need at least 2.5 GB free, on top of the 1.7 GB file already sitting in /tmp.
4. Check the copy is complete
sha256sum /tmp/rbb-rhel8-repo-20261117.tar.gz
ls /tmp/extra-rpms | wc -l
The code must match B's exactly, and the count should be 17.
5. Move the files into place and give them to root
mkdir -p /root/releases/2026.11
mv /tmp/rbb-rhel8-repo-20261117.tar.gz /root/releases/2026.11/
mv /tmp/extra-rpms /root/releases/2026.11/
mv /tmp/os_patch_wizard_1.5.2.sh /root/
chown -R root:root /root/releases/2026.11 /root/os_patch_wizard_1.5.2.sh
chmod +x /root/os_patch_wizard_1.5.2.sh
ls -l /root/releases/2026.11
Make sure the tar file is the only archive in that folder, so the wizard picks it automatically.
Back on B, as root (only after step 4 matched)
6. Clean up
rm /var/www/html/rbb-rhel8-repo-20261117.tar.gz
rm -rf /tmp/extra-rpms /tmp/os_patch_wizard_1.5.2.sh
df -h /
On A, as root: run the wizard
7. Start it
/root/os_patch_wizard_1.5.2.sh
8. Check Settings (C)
- Release:
2026.11. If it asks "Clear the old progress ticks?", answer y. - Repo: it should show
rhel-8-10-nov2026. - Environment: A's inventory.
- Servers to patch:
rbb-ref-a-*:!rbb-ref-a-vlim06
9. Steps 1 to 3
- Step 2: answer y to "Copy them into the repo?" for the extra RPMs. Answer y if it offers to switch off
databridgev6. - Step 3: every server should come back OK.
10. Step 4: patch
Patch the test server (vlim02) first, then the rest, then the master last. After the master is patched, reboot it and check the kernel:
uname -r # should be 4.18.0-553.185.1.el8_10.x86_64
11. Steps 5, 7 to 12, then Step 13
When you see "Run Step 13 anyway?", answer y.
12. Final checks
Confirm databridgev6 is still enabled = 0 on the A servers, then run the Nessus rescans on both sides.
Before Step 4, confirm with your senior whether A is the Active site right now and whether the services need to move first.
======================Prev Run Info======================================
All 17 files are there, including python3-unbound-1.16.2-5.14.el8_10.4. You don't need to add anything. I was wrong to say there should be 18: the 17 already include python3-unbound. In my earlier steps, wherever I said "should be 18", read 17.
The set is complete:
- kernel: 8 files, all
553.185.1 - vim: 4 files, all
8.0.1763-34 - unbound: 3 files, all
1.16.2-5.14.el8_10.4 - singles: compat-openssl10, libtiff, rhc
About the file permissions
The files are rw-r-----, so other users and Apache can't read them. That's fine: in Step 2 on A, the wizard sets every RPM in the repo to 644 (readable by Apache).
Next: the space check
Run the safe deletes, then look inside the three big folders:
rm -rf /var/www/html/os-patch.old-20261113-0538
rm -rf /var/www/html/rhel-8-10-rhel-8-10-oct2026
dnf clean all
df -h /
du -sh /root/releases/* | sort -h
du -sh /tmp/* 2>/dev/null | sort -h | tail -10
du -sh /var/log/* | sort -h | tail -10
Send me the output and I'll tell you what else is safe to remove. After that, carry on from step 3 of the A-side steps: copy extra-rpms to /tmp and hand it to dbv6admin.
===============================================
Don't make a new tar file on B. The disk is 98% full, with only 1.3 GB free, and a new tar would need about 1.7 GB.
You don't need one anyway. The file you used for B is still there: rbb-rhel8-repo-20261117.tar.gz (1.7 GB). The wizard moved it into /var/www/html during Step 2. Copy that file to A, together with the extra RPMs folder. A then gets exactly what B got, and nothing new has to be written on B.
1. Check the extra RPMs are still in the release folder
sudo ls /root/releases/2026.11/extra-rpms
You should see the 17 extra RPMs plus python3-unbound. If that folder is missing or empty, stop and tell me before deleting anything.
2. See what is using the space
sudo du -sh /var/www/html/* /root/* /var/cache/dnf /var/log /tmp 2>/dev/null | sort -h | tail -15
3. Free space on B, safest first
| Item | What it is | Delete? |
|---|---|---|
/var/cache/dnf |
Yum's download cache | Yes, any time: sudo dnf clean all |
os-patch.old-20261113-0538 |
Backup of the old os-patch folder, from before it became a link |
Yes. Nothing uses it |
rhel-8-10-rhel-8-10-oct2026 |
Last month's repo (its name has the prefix twice) | Yes. B is patched from the Nov repo, and os-patch points to nov2026 |
rbb-rhel8-repo-20261117.tar.gz |
The artefact you'll copy to A | Yes, but only after it's safely copied to A |
baseos_redhat8-20261108T220631 |
An earlier download folder | Check with your senior first. I'm not sure what it was used for |
Leave these alone: base_orch_yum_repo-*, databridgev6_yum_repo-1.0.16, index.html, os-patch, and rhel-8-10-nov2026. Running systems still use them.
Commands for the safe deletes:
sudo dnf clean all
sudo rm -rf /var/www/html/os-patch.old-20261113-0538
sudo rm -rf /var/www/html/rhel-8-10-rhel-8-10-oct2026
df -h /
4. Copy to A, then delete the tar file from B
Copy the files across first:
scp /var/www/html/rbb-rhel8-repo-20261117.tar.gz root@<A-master>:/root/releases/2026.11/
scp -r /root/releases/2026.11/extra-rpms root@<A-master>:/root/releases/2026.11/
Then, on A, check the copy is complete:
ls -l /root/releases/2026.11/
The tar file's size should be exactly 1783448406, the same as on B. Only when it matches, delete it on B:
sudo rm /var/www/html/rbb-rhel8-repo-20261117.tar.gz
5. Check space on A before you start
df -h /
You need at least 4 GB free: 1.7 GB for the file and 2.1 GB for the unpacked repo. If A is just as full, clean it the same way first. Don't delete A's current repo (the one os-patch points to) until A is patched.
On A, the wizard picks up the date 20261117 from the file name and names the repo rhel-8-10-nov2026. In Step 2 it finds extra-rpms and asks "Copy them into the repo?". Answer y.
The wizard doesn't delete the tar file after unpacking it, which is why the disk filled up. I'll add a clean-up offer to Step 2 in the next version.
====================Update 1=================
A-side checklist, with what we learned on B built in
What B taught us
- Missing RPMs fail Step 3. python3-unbound was the example. Using B's finished repo means A gets every RPM, extras included.
- The unused
databridgev6repo breaks Step 2. Let the wizard switch it off. - The master must be rebooted after it's patched. Otherwise the checks report the old kernel.
- In Step 13, answer y to "Run Step 13 anyway?". The MOTD part doesn't touch repos.
1. On the B master: pack the finished repo
B's repo folder already holds the October RPMs, the 17 extras and python3-unbound. Pack it once and A gets exactly what B got.
du -sh /var/www/html/rhel-8-10-nov2026
cd /var/www/html
tar -czf /root/rhel-8-10-nov2026-20261118.tar.gz --exclude=repodata rhel-8-10-nov2026
The 20261118 in the name lets the wizard work out the repo name rhel-8-10-nov2026 by itself. The repo index (repodata) is left out on purpose, because the wizard builds a fresh one.
2. On the A master: check space and make the release folder
df -h /var/www/html /root
mkdir -p /root/releases/2026.11
You need free space of at least twice the du figure from step 1: once for the file, once for the unpacked repo.
3. Copy the file and the wizard to A
Copy them the same way you moved files before:
scp /root/rhel-8-10-nov2026-20261118.tar.gz root@<A-master>:/root/releases/2026.11/
scp os_patch_wizard_1.5.2.sh root@<A-master>:/root/
Put only this one file in that folder. Then the wizard picks it automatically. You don't need an extra-rpms folder this time.
4. Start the wizard on A and check Settings (C)
- Release:
2026.11. If it asks "Clear the old progress ticks?", answer y. - Repo: it should show
rhel-8-10-nov2026. - Environment: the A side's inventory.
- Servers to patch:
rbb-ref-a-*:!rbb-ref-a-vlim06(vlim06 is still unreachable).
5. Step 1, then Step 2
- If Step 2 reports
databridgev6can't be reached and offers to switch it off, answer y. - The leftover notes in the repo report (glusterfs, samba, sssd, kernel-abi-whitelists) are harmless, as on B.
6. Step 3: pre-checks
Every server should come back OK. If one fails, stop and send me the output.
7. Step 4: patch
Patch in this order:
- The test server (vlim02).
- The rest.
- The master, last.
Then reboot the master:
reboot
uname -r # after it comes back, should be 4.18.0-553.185.1.el8_10.x86_64
8. Steps 5 and 7 to 12 on A
Same as on B. Step 12 shows SKIP if nothing was removed.
9. Step 13: MOTD
When you see "Run Step 13 anyway?", answer y. At the end, the banner should show 553.185.1.
10. Final checks
ansible 'rbb-ref-a-*:!rbb-ref-a-vlim06' -i <A inventory> --ask-vault-pass -u dbv6admin -m shell -a "grep -A5 databridgev6 /etc/yum.repos.d/*.repo | grep enabled"
Each server should show enabled = 0. Then run the Nessus rescan on A.
One thing to confirm with your senior before Step 4: whether A is the Active site right now. If it is, ask whether the services need to move to B first. Step 3 also asks you to look at TUI/Architect on both sides.
If you'd like this as a document for the ticket, I can make it a short runbook.
Comments
Post a Comment