[LU-9291] LustreError: Cannot find agent for archive 0

• Status: Resolved
• Project: Lustre
• Type: Question/Request
• Priority: Minor
• Reporter: Michael Gilbert
• Assignee: WC Triage
• Resolution: Incomplete

Description

I'm having trouble with my Lustre HSM setup. I had it working just fine and then I must have done something because now none of the lfs hsm commands will work. I tried restarting everything but I can't get the posix copytool to receive any instructions. On the head node I am running this:

lhsmtool_posix --verbose --hsm-root /hsm --archive=1 /lustre-scratch/

I only have this one agent running and the MDS sees it:

 

[root@mds ~]# lctl get_param -n mdt.scratch-MDT0000.hsm.agents
uuid=42ce88a0-4df0-b0ce-bfc8-26eb46e045cf archive_id=1 requests=[current:0 ok:0 errors:0]

 

However, I receive the following messages on the MDS at regular intervals:

 

Apr  4 12:12:24 mds kernel: LustreError: 1411:0:(mdt_hsm_cdt_agent.c:339:mdt_hsm_agent_send()) scratch-MDT0000: Cannot find agent for archive 0: rc = -11
Apr  4 12:12:24 mds kernel: LustreError: 1411:0:(mdt_hsm_cdt_agent.c:339:mdt_hsm_agent_send()) Skipped 96 previous similar messages

 

 And sure enough, if I try to do any hsm archive commands, nothing happens. The command seems to return successful, but no actions have been performed and the copytool output doesn't show anything:

 

[root@head mkg52]# lfs hsm_state *
testing123: (0x00000000)
testing456: (0x00000000)
[root@head mkg52]# lfs hsm_archive *
[root@head mkg52]# lfs hsm_state *
testing123: (0x00000000)
testing456: (0x00000000)

 

I'm not sure what to make of this and can't figure out what's happening. Can anyone help?

Comments

[Quentin Bouget]

Hi,

Did you try to stop the copytool and restart it?