SA Bugzilla – Bug 5834
mass-check should retry connections to avoid intermittent client disconnects
Last modified: 2019-07-07 15:39:49 UTC
Every few days one of my mass-check --clients exits early due to a single failure to connect to the mass-check --server. It's probably due to timeouts inflicted by some clients on the other side of a congested WAN link monopolizing the server. I think the client should retry a few times with a fairly short (maybe 10 second) interval.
Hrm... I'm not sure if it's always a failure to reconnect or what... sometimes they just seem to die after requesting more messages, other times they exit with the verbose statistics (but seem to forget they're in client mode): status: 96% ham: 11 spam: 281 date: 2007-12-25 now: 2008-02-23 04:30:01 status: 98% ham: 11 spam: 285 date: 2007-12-25 now: 2008-02-23 04:30:02 status: 99% ham: 11 spam: 289 date: 2007-12-25 now: 2008-02-23 04:30:04 status: 100% ham: 11 spam: 291 date: 2007-12-25 now: 2008-02-23 04:30:05 status: completed run stage now: 2008-02-23 04:30:06 status: completed run in 163 seconds now: 2008-02-23 04:30:06 status: requesting 222 messages from server now: 2008-02-23 04:30:06 status: 97% ham: 83 spam: 2633 date: 2008-03-20 now: 2008-04-07 06:11:02 status: 98% ham: 84 spam: 2660 date: 2008-03-20 now: 2008-04-07 06:11:03 status: 99% ham: 85 spam: 2687 date: 2008-03-20 now: 2008-04-07 06:11:05 status: 100% ham: 86 spam: 2709 date: 2008-03-20 now: 2008-04-07 06:11:07 status: completed run stage now: 2008-04-07 06:11:07 status: completed run in 126 seconds now: 2008-04-07 06:11:07 network I/O: sent=5383729, received=991471 processed 79801 messages (failed to scan 0) in 3861 seconds (74.4 kmsgs/hr)
moving off 3.3.0
Closing old stale bug. I doubt anyone would even use client/server option these days.