[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: rsync fixups for secondary1



+1

1On 11/1/2008 3:31 AM, Matt Domsch wrote:
I'd like to do the following to prevent our rsync hosts such as
secondary1 from being overloaded.




diff --git a/configs/system/rsync.xinetd b/configs/system/rsync.xinetd
index 8976dd7..f5b5dbc 100644
--- a/configs/system/rsync.xinetd
+++ b/configs/system/rsync.xinetd
@@ -10,4 +10,7 @@ service rsync
          server          = /usr/bin/rsync
          server_args     = --daemon
          log_on_failure  += USERID
+       instances       = 10
+       per_source      = 2
+       max_load        = 20
  }
diff --git a/configs/system/rsyncd.conf.secondary b/configs/system/rsyncd.conf.secondary
index 6c12db1..9101e05 100644
--- a/configs/system/rsyncd.conf.secondary
+++ b/configs/system/rsyncd.conf.secondary
@@ -3,6 +3,7 @@ syslog facility = daemon
  dont compress = *.gz *.tgz *.zip *.z *.rpm *.deb *.bz2 *.iso
  use chroot = false
  transfer logging = false
+timeout = 600

  [fedora-secondary]
  path = /srv/pub/fedora-secondary




This limits the number of simultaneous rsync clients to 10, at most 2
per IP (I had to kill a client with 10 open connections), establishes
a max load above which it stops serving clients, and establishes a 10
minute I/O timeout (no I/O after 10 minutes, close the connection).

These numbers are completely arbitrary, but should decrease the load
seen on secondary1, which tonight was at

04:57:21 up 15 days, 11:42,  1 user,  load average: 95.35, 136.54, 99.04

Acks please.

Thanks,
Matt





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]