Download raw body.
On 2026-06-16 12:33, Atanas Vladimirov wrote: > On 2026-06-13 15:42, Atanas Vladimirov wrote: >> On 2026-06-13 13:16, Atanas Vladimirov wrote: >>> On 2026-06-13 12:13, Atanas Vladimirov wrote: >>>> On 2026-06-13 10:35, Atanas Vladimirov wrote: >>>>> On 2026-06-13 06:09, Lucas Raab wrote: >>>>>> On Fri, Jun 12, 2026 at 03:29:21PM +0300, Atanas Vladimirov wrote: >>>>>>> On 2026-06-12 04:05, Lucas Raab wrote: >>>>>>> > On Thu, Jun 11, 2026 at 12:23:56PM +0300, Atanas Vladimirov wrote: >>>>>>> > > On 2026-06-11 04:30, Lucas Raab wrote: >>>>>>> > > > On Wed, Jun 10, 2026 at 03:33:58PM +0300, Atanas Vladimirov wrote: >>>>>>> > > > > On 2026-06-10 13:24, Atanas Vladimirov wrote: >>>>>>> > > > > > Hey Team, >>>>>>> > > > > > >>>>>>> > > > > > Here is a patch to update the Telegraf to the recently released 1.39.0. >>>>>>> > > > > > >>>>>>> > > > > > This version includes and improvement to OpenNTPd input plugin. >>>>>>> > > > > > It now collects all stats, including the sensors, if any. >>>>>>> > > > > > [PR] - https://github.com/influxdata/telegraf/pull/18468 >>>>>>> > > > > > >>>>>>> > > > > > I've been running this version for the last couple of hours and no problems so far. >>>>>>> > > > > > Let me know what do you think. >>>>>>> > > > > > >>>>>>> > > > > > Please add me in CC because I'm not subscribed to the list. >>>>>>> > > > > > Best wishes, >>>>>>> > > > > > Atanas >>>>>>> > > > > >>>>>>> > > > > Resending because something went wrong with the formatting. >>>>>>> > > > > >>>>>>> > > > I came up with the same diff, but inputs.mem no longer works. I haven't >>>>>>> > > > dived in too deeply yet, but I suspect something changed in gopsutil. >>>>>>> > > >>>>>>> > > Hmm, what do you mean by "inputs.mem no longer works"? It works for >>>>>>> > > me, here >>>>>>> > > is the output from my machine >>>>>>> > >>>>>>> > 2026-06-12T01:04:02Z E! [inputs.mem] Error in plugin: error getting >>>>>>> > virtual memory info: short syscall ret 104 bytes >>>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> I suppose that you're running -current, or not? >>>>>>> The server I'm testing runs Jun 1 snapshot if it matters at all. >>>>>>> >>>>>>> ``` >>>>>>> OpenBSD 7.9-current (GENERIC.MP) #543: Mon Jun 1 11:45:38 MDT >>>>>>> 2026 >>>>>>> >>>>>>> deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP >>>>>>> real mem = 34154913792 (32572MB) >>>>>>> avail mem = 33091518464 (31558MB) >>>>>>> ``` >>>>>> >>>>>> Yep, I'm on -current. Example machine: >>>>>> OpenBSD 7.9-current (GENERIC.MP) #0: Thu Jun 11 14:39:21 MDT 2026 >>>>>> >>>>>> deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP >>>>> >>>>> Okay, I just upgraded my machine to - Build date: 1781288383 - Fri >>>>> Jun 12 18:19:43 UTC 2026 >>>>> and got the same error like you :) >>>>> >>>>> So something has changed between Jun 1 - Jun 11. >>>> >>>> I think I found it - >>>> https://marc.info/?l=openbsd-cvs&m=178104984860610&w=2 >>>> >>>> Now I wonder if the following commit is the right fix for the >>>> gopsutil: >>>> https://github.com/shirou/gopsutil/commit/1d3af56f58e7ddcf3438bbcf09b632c3e7296b4f >>>> >>>> I'll create a gopsutil PR if you confirm that. >>> >>> I managed to build the telegraf against the patched gopsutil (the >>> mentioned commit) and >>> it now works as it should :) >>> >>> Now the question - is there a way to patch the OpenBSD port to >>> include this change until the >>> gopsutil accept the change, release a new version, and then the >>> telegraf release a new version with >>> the newer gopsutil? >>> >>> By the way, I believe that the current Telegraf version (1.38.3) from >>> the ports should be broken >>> since the Bob Beck's Jun 10 change (Remove the buffer flipper), >>> right? >>> (I might try that later today/tomorrow) >> >> Yes, the 1.38.3 has the same problem: >> >> ``` >> [hodor]/tmp$ doas ./telegraf --test --input-filter mem >> 2026-06-13T12:39:44Z W! Strict environment variable handling is the >> new default starting with v1.38.0! If your configuration does not work >> with strict handling please explicitly add the >> --non-strict-env-handling flag to switch to the previous behavior! >> 2026-06-13T12:39:44Z I! Loading config: /etc/telegraf/telegraf.conf >> 2026-06-13T12:39:44Z I! Starting Telegraf 1.38.3 brought to you by >> InfluxData the makers of InfluxDB >> 2026-06-13T12:39:44Z I! Available plugins: 243 inputs, 9 aggregators, >> 35 processors, 26 parsers, 68 outputs, 6 secret-stores >> 2026-06-13T12:39:44Z I! Loaded inputs: mem >> 2026-06-13T12:39:44Z I! Loaded aggregators: >> 2026-06-13T12:39:44Z I! Loaded processors: >> 2026-06-13T12:39:44Z I! Loaded secretstores: >> 2026-06-13T12:39:44Z W! Outputs are not used in testing mode! >> 2026-06-13T12:39:44Z I! Tags enabled: host=hodor.bsdbg.net >> 2026-06-13T12:39:44Z W! [agent] The default value of >> 'skip_processors_after_aggregators' will change to 'true' with >> Telegraf v1.40.0! If you need the current default behavior, please >> explicitly set the option to 'false'! >> 2026-06-13T12:39:44Z E! [inputs.mem] Error in plugin: error getting >> virtual memory info: short syscall ret 104 bytes >> 2026-06-13T12:39:44Z E! [telegraf] Error running agent: input plugins >> recorded 1 errors >> ``` > > Hello, > > Any thoughts if the > https://github.com/shirou/gopsutil/commit/1d3af56f58e7ddcf3438bbcf09b632c3e7296b4f > is the right thing to do? I can send it as PR to gopsutil if you say > so. Okay, https://github.com/shirou/gopsutil/pull/2104 was merged :) Now, we have to wait for a the new release of shirou/gopsutil with this commit included. Then, a new release of Telegraf that will depend on the newer shirou/gopsutil version :)