perfsonar-user - Re: [perfsonar-user] Failing throughput tests
Subject: perfSONAR User Q&A and Other Discussion
List archive
- From: Mark Feit <>
- To: "Smith, Sebastian" <>, "" <>
- Subject: Re: [perfsonar-user] Failing throughput tests
- Date: Tue, 17 Dec 2024 16:34:23 +0000
- Arc-authentication-results: i=1; mx.microsoft.com 1; spf=pass smtp.mailfrom=internet2.edu; dmarc=pass action=none header.from=internet2.edu; dkim=pass header.d=internet2.edu; arc=none
- Arc-message-signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=microsoft.com; s=arcselector10001; h=From:Date:Subject:Message-ID:Content-Type:MIME-Version:X-MS-Exchange-AntiSpam-MessageData-ChunkCount:X-MS-Exchange-AntiSpam-MessageData-0:X-MS-Exchange-AntiSpam-MessageData-1; bh=9R++AkMaezSU9/T8/ltUMCCvfr/27sdfTWJA4BGfw0c=; b=qsXK1ZzuzNb8RqXlG/F1syF4VZmULRc/NcyKljs63pRlGP3vkfU5XBgkYPYRrpbyjJT4Tw/m6o5feFS5cMwfX5USD+A1Zugh4RZ7tyBSmIbxc+9JsLPKNANBZUnI8SzmxQcL9MZvBc4QSmYARUslJlOmy0uYcfnvN79fTJNTbqt0LQ0s/heOvY8w6gkXGByrCqTqdtnhm6zJ0Lnso+vhBiJ/AfDiyzKjVxSnkawxqB3cTbg4dB3gI1KZ4qOBRx462UcQbBQ48+vgSgbONY7MpKL3WNNR2Qaqx4Je5xpf9VohNtbnn8WStADtrp0nr8VCtecv0w/u147wssD2BrUl+g==
- Arc-seal: i=1; a=rsa-sha256; s=arcselector10001; d=microsoft.com; cv=none; b=ifGhL4N6MmK9FLTiQhiTmQRXT7rTTORzIqkZzjLCka2C0vaqglABRRTidjcYyt3idZpY47uYN55ebZYwLPFY6/2zq90YfaZ0xgmfYxWdQcp7dX616XT8dI+W1WBmyIUwNXTYqFTR0EZrRs8Ro24esm390ejU0OrPhTHbtKkaHXDtdJyMvz2u/3oyojoQBf6A4lbOX6W0PBZEiyeBck4t1b9mgkOJWoohiRw56vOQyjR5vswkcBdfy+U+OpaWtXghXhDyFIWcMcM72En1F7QrASmR8oH2dsDxZtDiAiAIp7igrn5GF/CTdXuVD2kpCMBlJxZn76ZK0/b9N85WEUpjWA==
|
Smith, Sebastian writes:
When throughput tests fail, we’re seeing errors from `pscheduler troubleshoot`. Restarting logstash seems to clear the troubleshooter errors . Idle test.... 8 seconds... Pending, probably missed... Failed.
The likely case here is that the runner, the part of pScheduler that makes the measurements happen, has stopped running. pScheduler and Logstash don’t have any connection other than being installed on the same machine. If restarting Logstash makes the problem go away, check your test point’s system logs and make sure the OOM killer isn’t stopping processes because it’s short on memory. Some of the other things you’re seeing may be a side effect of that.
It’s also worth noting that an archive is not required if you’re just doing ad hoc measurements and looking at the results immediately. pScheduler will hang onto the results of measurements it does for 48 hours or until the archivings are completed or abandoned, whichever is later.
Dec 16 15:08:03 pdlperf2 archiver WARNING 58018: Failed to archive https://pdlperf2/pscheduler/tasks/e20eed17-e4fd-4f05-90de-b1ffe895c794/runs/d247fca1-3fcf-447a-aabb-ed1cbe4d099a to http: Failed to put result: 401: <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML 2.0//EN"> … Dec 16 15:08:03 pdlperf2 archiver WARNING <h1>Unauthorized</h1>
The archive where you’re sending the measurements needs to be configured to accept them and probably isn’t (see the docs: https://docs.perfsonar.net/multi_ma_install.html#authenticating-measurement-hosts). If your test points move around a lot and you have a list of subnets where they’ll be, Apache’s configuration can take those instead of individual IPs (e.g., Require ip 172.16.0.0/12 for the entire RFC1918 where the example you provided falls).
--Mark
|
- [perfsonar-user] Failing throughput tests, Smith, Sebastian, 12/16/2024
- Re: [perfsonar-user] Failing throughput tests, Mark Feit, 12/17/2024
Archive powered by MHonArc 2.6.24.