topic Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15 in ProLiant Servers (ML,DL,SL)

HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Thu, 25 Sep 2025 05:06:23 GMT

There is a issue again,. I have been following up this issue since v3.08 and indeed this was resolved in v3.10 & again I see this has started since v3.15 latest upgrade. I upgraded SPP release and even BIOS to U v3.60 but nothing helps and smad crashes very frquently. Again need to downgrade to v3.13 release. The same issue was seen on iLO 3.14 release as well.

----------------------------

smad[4127422]: [ERR ]: [AGX] Read size (0) < minimum packet size (20)
smad[4127422]: [ERR ]: Timeout for receiving OPEN_PDU after socket connected

# hponcfg -b
HPE Lights-Out Online Configuration utility
Version 6.0.0 Date 10/30/2022 (c) 2005,2022 Hewlett Packard Enterprise Development LP
Firmware Revision = 3.15 Device type = iLO 5 Driver name = hpilo
Rebooting the Management Processor...This takes upto 60 seconds.
The Management Processor has been rebooted.

----------------------------

Is there any solution available for this issue apart from manually rebooting iLO management processor ?

Parent Thread: Solved: Re: ILO5 - Proxmox - Debian 12 - Zabbix snmp monit... - Hewlett Packard Enterprise Community

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

BPSingh — Tue, 23 Sep 2025 01:13:43 GMT

Greetings!

If iLO downgrade works and the issue is seen with only certain iLO versions then the issue needs to be investigated.

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Mon, 29 Sep 2025 15:15:27 GMT

@BPSingh Thanks for confirmation. Is there any further updates from HPE on this SNMP fix as even now downgrading back in 1 case even does not fixes the issue and this time out occuring more frequently then. A new iLO 3.16 must be released with over SNMP fixes.

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Wed, 22 Oct 2025 13:01:33 GMT

iLO v3.16 was released recently even which upgrading still same SNMP failure issue exists. I see there was no SNMP fix mentioned in the v3.16 release notes.

Is there any news further on this case ?

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Fri, 26 Dec 2025 08:46:53 GMT

Recently iLO v3.17 was released and then amsd v4.4.0 was also released in parallel. I upgraded both and just after 48Hrs again same SMAD process starts to crash & NO SNMP data collection is seen under zabbix.

@BPSingh @HPE Any hope to fix this issue ?

This has become too much frequent and SNMP polling is effected.

smad[2213383]: [ERR ]: [AGX] Read size (0) < minimum packet size (20)
smad[2213383]: [ERR ]: Timeout for receiving OPEN_PDU after socket connected

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Mon, 19 Jan 2026 14:33:07 GMT

@BPSingh Is there anyone who can look into this thread and support for the fix of this issue ?

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

Azr_geek — Wed, 21 Jan 2026 10:47:03 GMT

Hello @SaquibAkhtar,

Reboot the iLO management processor (you already did that). Watch smad logs right after reboot. Check network connectivity between the host and iLO (link, IP, MTU). Make sure no firewall is blocking or dropping the connection.
Confirm the host’s HPE management packages (smad/AMSD, hpilo, hponcfg) and the OS kernel are versions supported for iLO 3.15. Mismatched drivers or old agent versions often cause problems.
If needed, stop or disable smad so it won’t crash the host (systemctl stop smad; systemctl disable smad), but only if you can lose the agent’s features. Or roll back iLO firmware to the last known good version (v3.13) in a maintenance window.
Collect logs for troubleshooting: Save smad logs, journalctl -u smad output, dmesg, hponcfg output, and any core dumps. Note exact versions of iLO, BIOS, SPP, OS, and kernel. These will help find the cause or open a support case.

Last option I would suggest is to create a HPE support ticket to check for the logs.

Regards,
Azr_geek

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

SaquibAkhtar — Wed, 21 Jan 2026 11:12:48 GMT

@Azr_geek -

Reboot the iLO management processor (you already did that). Watch smad logs right after reboot. Check network connectivity between the host and iLO (link, IP, MTU). Make sure no firewall is blocking or dropping the connection.>>> There is no FWR drop. After reboot of iLO as this the only temporary WA, then it works for some hours or days & then again smad.service crashes causing same SNMP timeout issue.
Confirm the host’s HPE management packages (smad/AMSD, hpilo, hponcfg) and the OS kernel are versions supported for iLO 3.15. Mismatched drivers or old agent versions often cause problems.

>>> I have upgraded to iLO 3.17, amsd package to v4.4.0 release , OS = RHEL 9.6 , RHEL 9.7 but even though nothing got fixed.

If needed, stop or disable smad so it won’t crash the host (systemctl stop smad; systemctl disable smad), but only if you can lose the agent’s features. Or roll back iLO firmware to the last known good version (v3.13) in a maintenance window.

>>> If I stop smad then SNMP timeout occurs and the only solution is to do: hponcfg -b // iLO graceful reboot.

Collect logs for troubleshooting: Save smad logs, journalctl -u smad output, dmesg, hponcfg output, and any core dumps. Note exact versions of iLO, BIOS, SPP, OS, and kernel. These will help find the cause or open a support case.

>>> So finally a support case needed ? I was looking help on this from Community end if somehow possible. The most weird thing is there is no such issue with HPE Gen9 Systems on iLO v4 running same RHEL 9.6 OS.

If you see below error then smad crashes stating: [AGX] Read size (0) < minimum packet size (20) resulting into no SNMP data collection by Zabbix Monitoring tool.

smad[2212271]: [INFO ]: BMC device is /dev/hpilo/d0ccb1
smad[2212271]: [NOTICE]: iLO5: 3.17 (2025/12/2)
smad[2212271]: [INFO ]: Reset iLO AgentX session
smad[2212271]: [INFO ]: [CHIF] Clear out buffer
smad[2212271]: [INFO ]: [CHIF] Close previous AgentX session if any
smad[2212271]: [INFO ]: [CHIF] AgentX session is ready to use
smad[2212271]: [INFO ]: Open iLO AgentX session in Regular Mode
smad[2212271]: [NOTICE]: (size = 12 cmd = 93 seq = errcode=0
smad[2212271]: [ERR ]: [AGX] Read size (0) < minimum packet size (20)
smad[2212271]: [ERR ]: Timeout for receiving OPEN_PDU after socket connected

Re: HPE Gen10 AMSD on Linux smad crash, iLO v3.15

Azr_geek — Wed, 21 Jan 2026 11:23:50 GMT

@SaquibAkhtar

Logs and info to collect for troubleshooting:
- journalctl -u smad -n 200 (or full smad logs), dmesg, /var/log/messages or syslog
- hponcfg -w output and hpilo/hponcfg versions
- Any core dumps, timestamps of crashes, and whether the crash happens right after iLO reboot
- Network checks (ping, tcpdump on iLO management interface) when the problem occurs
Next step: Open an HPE support case with the logs and exact version list (iLO firmware, host agent/AMSD, SPP, BIOS, OS/kernel). This looks like a firmware/agent compatibility or iLO protocol regression and HPE may provide a hotfix or patch.