Bug: gvmd aborts (crashes) on duplicate-key constraint violation during NVT rebuild instead of handling gracefully

[ShreyaTrivedi1209]

This is an issue and not a support question which should be asked at https://forum.greenbone.net/?

• This is an issue and not a support question

Is there an existing issue for this?

• I have searched the existing issues

Current Behavior

gvmd version: 26.24.0 (DB revision 273)
OS: Ubuntu 24.04.4 LTS, built from source
Feed version: 202607020814

During a full NVT rebuild triggered by a feed hash mismatch, update_nvts_from_osp_vts inserts into the nvts_rebuild staging table and hits a duplicate key violation on OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375:

ERROR: duplicate key value violates unique constraint "nvts_rebuild_uuid_key"
DETAIL: Key (uuid)=(1.3.6.1.4.1.25623.1.1.1.1.2003.375) already exists.

This SQL error is unhandled and causes the entire gvmd process to receive an Aborted signal and crash, which then restarts, re-attempts the same rebuild, and crashes again in an infinite loop — the feed can never finish syncing.

The duplicate OID appears to originate from the same OID existing in both a NASL plugin (/var/lib/openvas/plugins/2023/debian/deb_375.nasl) and a Notus advisory (/var/lib/notus/advisories/debian.notus) simultaneously.

Expected Behavior

Expected: gvmd should catch this constraint violation, log a warning, and skip/deduplicate the conflicting row rather than crashing the whole daemon.

md main:MESSAGE:2026-07-06 10h55.51 utc:938079: Rebuilding all NVTs because of a hash value mismatch
md manage:WARNING:2026-07-06 10h57.11 utc:938079: sql_exec_internal: constraint violation: ERROR: duplicate key value violates unique constraint "nvts_rebuild_uuid_key"
DETAIL: Key (uuid)=(1.3.6.1.4.1.25623.1.1.1.1.2003.375) already exists.
md manage:WARNING:2026-07-06 10h57.11 utc:938079: sql_exec_internal: SQL: INSERT into nvts_rebuild (oid, name, summary, insight, affected, impact, cve, tag, category, family, cvss_base, creation_time, modification_time, uuid, solution_type, solution_method, solution, detection, qod, qod_type) VALUES ('1.3.6.1.4.1.25623.1.1.1.1.2003.375', 'Debian: Security Advisory (DSA-375)', 'The remote host is missing an update for the Debian ''node'' package(s) announced via the DSA-375 advisory.', 'Morgan alias SM6TKY discovered and fixed several security related problems in LinuxNode, an Amateur Packet Radio Node program. The buffer overflow he discovered can be used to gain unauthorised root access and can be remotely triggered.

For the stable distribution (woody) this problem has been fixed in version 0.3.0a-2woody1.

For the unstable distribution (sid) this problem has been fixed in version 0.3.2-1.

We recommend that you upgrade your node packages immediately.', '''node'' package(s) on Debian 3.0.', '', 'CVE-2003-0707, CVE-2003-0708', 'cvss_base_vector=AV:N/AC:L/Au:N/C:P/I:P/A:P', 3, 'Debian Local Security Checks', 7.5, 1678280204, 1706776332, '1.3.6.1.4.1.25623.1.1.1.1.2003.375', 'VendorFix', '', 'Please install the updated package(s).', 'Checks if a vulnerable package version is present on the target host.', 97, 'package');
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x141510) [0x5a9697abd510]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(+0x45330) [0x725b5ba45330]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(pthread_kill+0x11c) [0x725b5ba9eb2c]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(gsignal+0x1e) [0x725b5ba4527e]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(abort+0xdf) [0x725b5ba288ff]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x5c2f9) [0x5a96979d82f9]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x13fa65) [0x5a9697abba65]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x13e74e) [0x5a9697aba74e]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x13ef47) [0x5a9697abaf47]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x142abc) [0x5a9697abeabc]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x1060ca) [0x5a9697a820ca]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0xa865e) [0x5a9697a2465e]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x1421f7) [0x5a9697abe1f7]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x1450cd) [0x5a9697ac10cd]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(+0x2a1ca) [0x725b5ba2a1ca]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0x8b) [0x725b5ba2a28b]
md main:MESSAGE:2026-07-06 10h57.11 utc:938079: BACKTRACE: gvmd: OSP: Updatin(+0x5c365) [0x5a96979d8365]
md manage:MESSAGE:2026-07-06 10h57.11 utc:938079: Received Aborted signal

root@IAAHMCPU298:~#

Steps To Reproduce

1. Have a gvmd database where nvts_feed_version in the meta table is 0 (or otherwise behind the scanner's reported feed version) while the nvts table already contains data from a prior sync.
2. Start gvmd with ospd-openvas reporting a current feed version (e.g. 202607020814).
3. gvmd detects the mismatch and begins update_nvts_from_osp_vts, successfully inserting ~225,000 new VTs.
4. It then performs a hash comparison, finds a mismatch, and triggers "Rebuilding all NVTs."
5. During that second rebuild pass, the INSERT into nvts_rebuild fails on OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375 with a duplicate key violation, because this OID exists in both a NASL plugin file and a Notus advisory JSON file simultaneously.
6. gvmd aborts and restarts, repeating the same failure indefinitely.

Operating System

Ubuntu 24.04.4 LTS, built from source

Version

26.24.0 (DB revision 273)
feed version: 202607020814

Anything else?

No response

[cfi-gb]

gvmd version: 26.24.0 (DB revision 273)

If you check https://github.com/greenbone/gvmd/releases you are lacking behind multiple versions. Please make sure to update all components (not only gvmd) to the most recent versions before raising any issues.

[ShreyaTrivedi1209]

gvmd version: 26.24.0 (DB revision 273)

If you check https://github.com/greenbone/gvmd/releases you are lacking behind multiple versions. Please make sure to update all components (not only gvmd) to the most recent versions before raising any issues.

Thanks for pointing that out — you're right, we're on 26.24.0 and hadn't checked recent releases closely. I see 26.28.0 includes "Improve error handling when reading vts from openvasd" (#2929), which looks like it could be directly relevant to this crash. We'll update to the latest release (26.29.0) and gvm-libs/openvas-scanner/ospd-openvas to their current versions, then retest a full NVT rebuild to see if the issue persists. Will report back with results.

[cfi-gb]

Could be also a duplicate of #2480

[cfi-gb]

https://forum.greenbone.net/t/duplicate-oid-1-3-6-1-4-1-25623-1-1-1-1-2003-375-present-in-both-nasl-and-notus-feed-simultaneously-feed-version-202607020814/22397 also indicates that this is a setup issue or similar (There are no two variants, a NASL and a notus based for OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375 in a feed at the same time).

Maybe you are syncing a 21.04 feed and using a notus based feed in top?

[ShreyaTrivedi1209]

https://forum.greenbone.net/t/duplicate-oid-1-3-6-1-4-1-25623-1-1-1-1-2003-375-present-in-both-nasl-and-notus-feed-simultaneously-feed-version-202607020814/22397 also indicates that this is a setup issue or similar (There are no two variants, a NASL and a notus based for OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375 in a feed at the same time).

Maybe you are syncing a 21.04 feed and using a notus based feed in top?

We sync the feed via direct rsync from rsync://feed.community.greenbone.net, not greenbone-feed-sync. I suspect our rsync command may be missing --delete, meaning retired NASL files (like this Debian DSA-375 check, likely superseded by its Notus equivalent) never get cleaned up locally after Greenbone removes them upstream — leaving a stale legacy file to collide with the new Notus advisory carrying the same OID. We'll check our sync command and re-sync with proper deletion of stale files, then confirm whether that resolves it

[ShreyaTrivedi1209]

https://forum.greenbone.net/t/duplicate-oid-1-3-6-1-4-1-25623-1-1-1-1-2003-375-present-in-both-nasl-and-notus-feed-simultaneously-feed-version-202607020814/22397 also indicates that this is a setup issue or similar (There are no two variants, a NASL and a notus based for OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375 in a feed at the same time).
Maybe you are syncing a 21.04 feed and using a notus based feed in top?

We sync the feed via direct rsync from rsync://feed.community.greenbone.net, not greenbone-feed-sync. I suspect our rsync command may be missing --delete, meaning retired NASL files (like this Debian DSA-375 check, likely superseded by its Notus equivalent) never get cleaned up locally after Greenbone removes them upstream — leaving a stale legacy file to collide with the new Notus advisory carrying the same OID. We'll check our sync command and re-sync with proper deletion of stale files, then confirm whether that resolves it

we upgraded to gvmd 26.29.0 and re-synced using the official greenbone-feed-sync 25.3.0 tool (previously used manual per-module rsync, now abandoned in favor of the official tool). Both NASL and Notus data are freshly synced from today. The crash on OID 1.3.6.1.4.1.25623.1.1.1.1.2003.375 still occurs identically. We also checked file timestamps directly: /var/lib/openvas/plugins/2023/debian/deb_375.nasl and /var/lib/notus/advisories/debian.notus both carry recent, distinct upstream modification times (preserved via rsync), suggesting this OID genuinely exists as both a NASL check and a Notus advisory in the current feed snapshot (202607020814) on Greenbone's server — not a local sync artifact. Could you confirm on your end whether this specific OID is duplicated in the current feed?

[cfi-gb]

For receiving such support i would suggest to use the community forums:

https://forum.greenbone.net/t/duplicate-oid-1-3-6-1-4-1-25623-1-1-1-1-2003-375-present-in-both-nasl-and-notus-feed-simultaneously-feed-version-202607020814/22397/2

as suggested by the filled out template:

This is an issue and not a support question which should be asked at https://forum.greenbone.net/?

[ShreyaTrivedi1209]

For receiving such support i would suggest to use the community forums:

https://forum.greenbone.net/t/duplicate-oid-1-3-6-1-4-1-25623-1-1-1-1-2003-375-present-in-both-nasl-and-notus-feed-simultaneously-feed-version-202607020814/22397/2

as suggested by the filled out template:

This is an issue and not a support question which should be asked at https://forum.greenbone.net/?

The Issue was raised there as well, just waiting on a reply on that.