Date: Fri, 27 Sep 2024 13:45:41 +0100 From: Frank Leonhardt <freebsd-doc@fjl.co.uk> To: questions@freebsd.org Subject: Re: Samba hogs syslogd and fails Message-ID: <981a22c7936ef27ae5e43b42d9c27956@fjl.co.uk> In-Reply-To: <e3bc196e-552c-4ea3-8cac-5d0dea68ee0f@netfence.it> References: <d3035dee-c39f-4c6a-aa23-3caf2d98f6bf@netfence.it> <fd0e5c8ac9c4722e56df08d66aaa19bb@fjl.co.uk> <e3bc196e-552c-4ea3-8cac-5d0dea68ee0f@netfence.it>
next in thread | previous in thread | raw e-mail | index | archive | help
On 2024-09-27 12:47, Andrea Venturoli wrote: > On 9/27/24 12:38, Frank Leonhardt wrote: > >> As no one else who knows better has answered... >> ... >> If you didn't know this already, I hope it helps. > > Thanks. > I had dealt with a similar problem long ago, so, more or less, I knew, > but a refresh was useful. > > I fully understand that this will only delay the problem, not solve it. > I guess "solving" it should be a Samba problem, not an OS one, unless > there's a way to force the syslog API to stop and wait when the buffers > are exahusted. > > I'll try anyway to increase these values, but the real problem is > vfs_full_audit won't tell when this happen, so you have to know you > lost logs... and usually you don't :( Here's a crazy idea. syslogd does a DNS lookup to log the originating hostname. Try running it with the -n option to disable this if it can't resolve your hostname - DNS timeouts are a bitch. Would it be possible to clear the O_NONBLOCK flag on the syslog socket? I would have thought so but I'm not sure how (other than recompiling syslogd). -- ------ 25-Sept-24 My apologies to everyone who I appear to have ignored for the last few years. A procmail script was misfiling some replies to Questions to the wrong folder.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?981a22c7936ef27ae5e43b42d9c27956>