Hello, I successfully use tinc for one year in a configuration with several subnets (about 5-7 subnets) and one main server. Sometimes tinc dies with the error message: Error while waiting for input: Bad file descriptor If it happens, this happens immediately, when one subnet has disconnected. tinc (on the main server) tries then to disconnect all currently connected subnets and while disconnecting the last active subnet the above error message appears. It seems, that subnets that connect and disconnect often (dial-in subnets), can force this problem faster, but not I cannot reproduce it with a known sequence of commands or connections. After restarting tinc on the main server all subnets reconnects after some time and all is fine. I use the tinc version of debian woody (this means 1.0-pre4 with some changes of the cvs-version). What can be done to resolve this problem? Flynn Tinc: Discussion list about the tinc VPN daemon Archive: http://mail.nl.linux.org/lists/ Tinc site: http://tinc.nl.linux.org/
Kostadin Galabov
2001-Dec-13 10:42 UTC
tinc dies after some days/weekes of normal activity
If I remember correct there was a fix for this problem in CVS version a long time ago. I never had it as I use pre3. Try to compile CVS version ----- Original Message ----- From: "Flynn Marquardt" <flynn@onestepahead.de> To: <tinc@nl.linux.org> Sent: Monday, December 10, 2001 20:59 Subject: tinc dies after some days/weekes of normal activity> Hello, > > I successfully use tinc for one year in a configuration with several > subnets (about 5-7 subnets) and one main server. Sometimes tinc dies > with the error message: > > Error while waiting for input: Bad file descriptor > > If it happens, this happens immediately, when one subnet has > disconnected. tinc (on the main server) tries then to disconnect all > currently connected subnets and while disconnecting the last active > subnet the above error message appears. > > It seems, that subnets that connect and disconnect often (dial-in > subnets), can force this problem faster, but not I cannot reproduce it > with a known sequence of commands or connections. > > After restarting tinc on the main server all subnets reconnects after > some time and all is fine. > > I use the tinc version of debian woody (this means 1.0-pre4 with some > changes of the cvs-version). > > What can be done to resolve this problem? > > Flynn > > Tinc: Discussion list about the tinc VPN daemon > Archive: http://mail.nl.linux.org/lists/ > Tinc site: http://tinc.nl.linux.org/ > >Tinc: Discussion list about the tinc VPN daemon Archive: http://mail.nl.linux.org/lists/ Tinc site: http://tinc.nl.linux.org/
On Thu, Dec 13, 2001 at 11:42:03AM +0200, Kostadin Galabov wrote:> If I remember correct there was a fix for this problem in CVS version a long > time ago. I never had it as I use pre3. Try to compile CVS versionI do not remember this problem. Do you remember exactly how long ago it was fixed? -- Met vriendelijke groet / with kind regards, Guus Sliepen <guus@sliepen.warande.net> -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 232 bytes Desc: not available Url : http://brouwer.uvt.nl/pipermail/tinc/attachments/20011213/5f52de48/attachment.pgp
On Mon, Dec 10, 2001 at 07:59:20PM +0100, Flynn Marquardt wrote:> I successfully use tinc for one year in a configuration with several > subnets (about 5-7 subnets) and one main server. Sometimes tinc dies > with the error message: > > Error while waiting for input: Bad file descriptor[...] Can you tell me what kernel version you are using and could you provide me with all syslog messages generated by tinc? -- Met vriendelijke groet / with kind regards, Guus Sliepen <guus@sliepen.warande.net> -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 232 bytes Desc: not available Url : http://brouwer.uvt.nl/pipermail/tinc/attachments/20011213/dc283f78/attachment.pgp
On Thu, Dec 13, 2001 at 12:22:31PM +0100, Guus Sliepen wrote:> On Thu, Dec 13, 2001 at 11:42:03AM +0200, Kostadin Galabov wrote: > > > If I remember correct there was a fix for this problem in CVS version a long > > time ago. I never had it as I use pre3. Try to compile CVS version > > I do not remember this problem. Do you remember exactly how long ago it > was fixed?This sounds exaclty like a problem we had running pre1. The problem was that socket filedescriptors were not being closed. After tinc had run long enough and enough connections had come and gone the maximum file descriptor limit was reached and the bad filedescriptor messages started appearing. It was fixed in pre2, as I recall. - Jamie Tinc: Discussion list about the tinc VPN daemon Archive: http://mail.nl.linux.org/lists/ Tinc site: http://tinc.nl.linux.org/