From owner-freebsd-questions@freebsd.org Sat Sep 2 22:36:17 2017 Return-Path: Delivered-To: freebsd-questions@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 07EF5E07E48 for ; Sat, 2 Sep 2017 22:36:17 +0000 (UTC) (envelope-from zi@FreeBSD.org) Received: from exodus.zi0r.com (exodus.zi0r.com [71.179.14.195]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (Client CN "exodus.zi0r.com", Issuer "Gandi Standard SSL CA 2" (not verified)) by mx1.freebsd.org (Postfix) with ESMTPS id CAB5672EA6; Sat, 2 Sep 2017 22:36:16 +0000 (UTC) (envelope-from zi@FreeBSD.org) Received: from exodus.zi0r.com (localhost [127.0.0.1]) by exodus.zi0r.com (Postfix) with ESMTP id 29835BD3B3; Sat, 2 Sep 2017 18:36:15 -0400 (EDT) X-Virus-Scanned: amavisd-new at zi0r.com Received: from exodus.zi0r.com ([127.0.0.1]) by exodus.zi0r.com (exodus.zi0r.com [127.0.0.1]) (amavisd-new, port 10026) with LMTP id Fn-gSLqGsJ6l; Sat, 2 Sep 2017 18:36:14 -0400 (EDT) Received: from exodus.zi0r.com (syn.zi0r.com [IPv6:2001:470:8:d58:aba:ca:daba:10]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by exodus.zi0r.com (Postfix) with ESMTPSA id 88668BD3B2; Sat, 2 Sep 2017 18:36:14 -0400 (EDT) Date: Sat, 2 Sep 2017 18:36:12 -0400 From: Ryan Steinmetz To: Rafal Lukawiecki Cc: freebsd-questions@freebsd.org Subject: Re: pkg.freebsd.org mirror having issues? Message-ID: <20170902223612.GA66389@exodus.zi0r.com> References: <20170902212006.GA70826@exodus.zi0r.com> <54C54D41-E4D0-4E20-A2F2-498AD0D0187E@rafal.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Disposition: inline In-Reply-To: <54C54D41-E4D0-4E20-A2F2-498AD0D0187E@rafal.net> User-Agent: Mutt/1.8.3 (2017-05-23) X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Sep 2017 22:36:17 -0000 On (09/02/17 22:22), Rafal Lukawiecki wrote: >Ryan, > >I am glad to hear there is a reason for it. I am still awaiting a server to finish building, it has been some 25 minutes, looks like 2 min timeouts between files. :) Looks like I misspoke. This actually didn't have anything to do with my testing, but rather another mirror that was having issues. The 2 minute delay you were experiencing was caused by pkg waiting for a timeout on another mirror before trying one that is responding. I've removing the offending mirror for now and will investigate a more automatic solution in the near future. -r > >I will terminate this server instance and start over. I will report back here as soon as I can see it is faster. > >BTW, are you/others on an IRC/Freenode to let you know of such issues more directly? > >Thank you and regards from Ireland, >Rafal > >> On 2 Sep 2017, at 22:20, Ryan Steinmetz wrote: >> >> >>> As a new FreeBSD user I wonder if it is a frequent problem. I am suddenly experiencing very slow access to the local mirrors of pkg.freebsd.org. Even a small file takes almost a minute. Updating FreeBSD repository catalogue (when using pkg) takes about 5-6 minutes. I have tried using several different machines from different IPs, all within AWS eu-west (Ireland). It was working well all week. >>> The local mirror seems to resolve to dualstack.nonssl.global.fastly.net (151.101.60.204) at the moment. Pingtime is not bad, 9ms, but their server takes almost a minute before it starts serving files. Once the file starts, it comes fast. I assume it is some sort of a server issue at their end at Fastly, perhaps they have a dead machine in the pool which makes it intermittent. >>> As I am building an automated server init, this means that instead of having a server provisioned in <5 min, it can take 20-35 mins, by which time it is not even considered to exist by the load balancer. >>> Is this common? Should I start (somehow) specifying individual mirrors directly? Thank you very much for your wisdom. >> >> >> Rafal, >> >> This is not common and was caused by a configuration issue during some testing that I've been doing to (ironically) try to improve response time. >> >> Would you be able to try this once more? I've made an edit which should drastically change the experience for you. >> >> Thanks! >> -r >> >>> -- >>> Rafal Lukawiecki >>> Data Scientist Project Botticelli Ltd >> >> >> >> -- >> Ryan Steinmetz >> PGP: 9079 51A3 34EF 0CD4 F228 EDC6 1EF8 BA6B D028 46D7 > -- Ryan Steinmetz PGP: 9079 51A3 34EF 0CD4 F228 EDC6 1EF8 BA6B D028 46D7