GNU bug report logs - #48468
substitute server connection timeout

Previous Next

Full log

Message #37 received at 48468 <at> debbugs.gnu.org (full text, mbox):

From: Mathieu Othacehe <othacehe <at> gnu.org>
To: Ludovic Courtès <ludo <at> gnu.org>
Cc: 48468 <at> debbugs.gnu.org
Subject: Re: bug#48468: substitute server connection timeout
Date: Tue, 29 Jun 2021 18:49:26 +0200

Hey,

> But maybe I’m not looking at the right thing.
>
> Do you have evidence or a reproducer?

Yes, adding the following debug message here:

--8<---------------cut here---------------start------------->8---
--- a/guix/scripts/substitute.scm
+++ b/guix/scripts/substitute.scm
@@ -416,6 +416,7 @@ server certificates."
         ;; and retry.  We might also get 'bad-response or a similar
         ;; exception from (web response) later on, once we've sent the
         ;; request, or a ERROR/INVALID-SESSION from GnuTLS.
+        (pk key args)
         (if (or (and (eq? key 'system-error)
                      (= EPIPE (system-error-errno `(,key ,@args))))
                 (and (eq? key 'gnutls-error)
--8<---------------cut here---------------end--------------->8---

and using substitutes from a publish server without keep alive support
(before 0b8fa24), prints something like:

--8<---------------cut here---------------start------------->8---
downloading from http://192.168.1.51:8080/nar/gzip/lkmiyfzi1bq571yn4gypbcs5vn4fpma7-texlive-cm-51265 ...
 texlive-cm-51265                                                                                                                                                                                                                                      7.8MiB/s 00:00 | 1.7MiB transferred

substituting /gnu/store/l788x07ska5vffayz0gayv4hsx5flxal-module-import-compiled...

;;; (bad-header (read-header-line #<eof>))
downloading from http://192.168.1.51:8080/nar/gzip/l788x07ska5vffayz0gayv4hsx5flxal-module-import-compiled ...
 module-import-compiled                                                                                                                                                                                                                                 7.5MiB/s 00:00 | 85KiB transferred

substituting /gnu/store/1s1lrnxlkjwxshk5q2w97ig3clc6n06f-ruby-2.6.5...

;;; (bad-header (read-header-line #<eof>))
--8<---------------cut here---------------end--------------->8---

That's because the connection is closed by the publish server each time
a NAR is sent. This particular behaviour is fixed by:
https://issues.guix.gnu.org/48556.

I hoped that it would decrease the load of the Cuirass publish server
that was flooded by worker connections. While the situation is better,
there are still a lot of substitute timeout errors on Cuirass.

I have also observed those timeouts using the publish server behind
ci.guix.gnu.org. This makes me think that there's something else that is
preventing the server from honoring every connection request in less
than 5 seconds (the timeout duration).

Thanks,

Mathieu

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.