Folks, I think that my write_socket_data problems have been solved thanks to a good pointer by someone last week on the list. The system in question runs named in addition to smb, and I did not have a reverse-DNS file set up for the local, private-IP (192.168.x.x), domain. Since I did that on Friday afternoon these are my logs: [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/13" log.smb:[2000/11/13 16:01:39, 0] lib/util_sock.c:write_socket_data(540) [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/14" [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/15" [root@virtue samba]# Normally, this system has had about 10 write_socket_data errors per day since the errors started occuring, so the logs listed above show a dramatic improvement, and my guess would be that I probably won't see any more of it in the degree that I have seen it before. Thanks all, Lee Howard
Lee - This sounds like it might be the solution to the "Network is Busy" problem I've been having - can you describe the fix that worked for you? The "reverse DNS-file" comment in your post went past me <g>. Thanks, Steve McClary At 12:01 PM 11/15/2000 -0700, you wrote:>Folks, I think that my write_socket_data problems have been solved thanks >to a good pointer by someone last week on the list. > >The system in question runs named in addition to smb, and I did not have a >reverse-DNS file set up for the local, private-IP (192.168.x.x), domain. >Since I did that on Friday afternoon these are my logs: > >[root@virtue samba]# grep "write_socket_data" * | grep "2000/11/13" >log.smb:[2000/11/13 16:01:39, 0] lib/util_sock.c:write_socket_data(540) >[root@virtue samba]# grep "write_socket_data" * | grep "2000/11/14" >[root@virtue samba]# grep "write_socket_data" * | grep "2000/11/15" >[root@virtue samba]# > >Normally, this system has had about 10 write_socket_data errors per day >since the errors started occuring, so the logs listed above show a dramatic >improvement, and my guess would be that I probably won't see any more of it >in the degree that I have seen it before. > >Thanks all, > >Lee Howard
I can confirm that adding reverse address lookup seems to solve the problem. Since we added reverse DNS about a week ago, there has not been a single write_socket_data error message in the samba logs. Many thanks for this useful tip. Otto Giesenfeld> -----Original Message----- > From: Lee Howard [mailto:faxguy@deanox.com] > Sent: Wednesday, November 15, 2000 8:02 PM > To: samba@us5.samba.org > Subject: [SOLVED?] write_socket_data bug > > > Folks, I think that my write_socket_data problems have been > solved thanks > to a good pointer by someone last week on the list. > > The system in question runs named in addition to smb, and I > did not have a > reverse-DNS file set up for the local, private-IP > (192.168.x.x), domain. > Since I did that on Friday afternoon these are my logs: > > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/13" > log.smb:[2000/11/13 16:01:39, 0] > lib/util_sock.c:write_socket_data(540) > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/14" > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/15" > [root@virtue samba]# > > Normally, this system has had about 10 write_socket_data > errors per day > since the errors started occuring, so the logs listed above > show a dramatic > improvement, and my guess would be that I probably won't see > any more of it > in the degree that I have seen it before. > > Thanks all, > > Lee Howard-------------- next part -------------- HTML attachment scrubbed and removed
There seems to be relatively little information on DNS administration online. "Buy the book" is the most common advice. But Ask Mr DNS, http://www.acmebw.com/_vti_bin/shtml.exe/askmr.htm is a good place to start for online resources, and for the record, we have done reasonably well without books where I work. / Otto Giesenfeld -----Original Message----- From: Steve McClary [mailto:scm@mrwassoc.com] Sent: Tuesday, November 28, 2000 7:16 PM To: Otto Giesenfeld Cc: samba@samba.org Subject: RE: [SOLVED?] write_socket_data bug This sounds like exactly the solution to the problems I've been having - could someone describe the process of adding reverse DNS in simple terms for novices like me? Many thanks - the list has been quite helpful and I appreciate the gathered wisdom of all you subscribers. Steve McClary At 04:46 PM 11/28/2000 +0100, Otto Giesenfeld wrote: I can confirm that adding reverse address lookup seems to solve the problem. Since we added reverse DNS about a week ago, there has not been a single write_socket_data error message in the samba logs. Many thanks for this useful tip. Otto Giesenfeld> -----Original Message----- > From: Lee Howard [mailto:faxguy@deanox.com] > Sent: Wednesday, November 15, 2000 8:02 PM > To: samba@us5.samba.org > Subject: [SOLVED?] write_socket_data bug > > > Folks, I think that my write_socket_data problems have been > solved thanks > to a good pointer by someone last week on the list. > > The system in question runs named in addition to smb, and I > did not have a > reverse-DNS file set up for the local, private-IP > (192.168.x.x), domain. > Since I did that on Friday afternoon these are my logs: > > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/13" > log.smb:[2000/11/13 16:01:39, 0] > lib/util_sock.c:write_socket_data(540) > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/14" > [root@virtue samba]# grep "write_socket_data" * | grep "2000/11/15" > [root@virtue samba]# > > Normally, this system has had about 10 write_socket_data > errors per day > since the errors started occuring, so the logs listed above > show a dramatic > improvement, and my guess would be that I probably won't see > any more of it > in the degree that I have seen it before. > > Thanks all, > > Lee Howard-------------- next part -------------- HTML attachment scrubbed and removed