I don't know how things have been going for many of you out there with Samba 2.2.2, but I'm having horrible results with the stability of Samba in a production environment. Using Samba 2.2.2 from the FreeBSD-stable ports collection: 1) Op locks have to be disabled. There are a horrible amount of errors that Samba logs in the machine specific logs. So many that eventually smbd crashes with a signal 6. 2) When operating in domain controller mode, we've noticed a couple of times now that samba just stops working as a domain controller. No errors in the smbd or nmbd logs, it just fails to authenticate logins, etc. Restarting Samba seems to fix the problem. 3) Some Windows 2K pro machines that are member machines (with trust accounts) seem to like to blue screen when they start up. Hard crash. Usually unpluging the network jack and waiting for the machine to boot fully, then pluging it back in seems to be fine. (probably a Windows and samba problem) We also see many machine logs that have errors as follows: [2002/01/07 14:37:18, 0] rpc_server/srv_pipe.c:api_rpcTNP(1204) api_rpcTNP: api_netlog_rpc: NET_SAMLOGON failed. Which most times doesn't seem to effect anything, but it's rather annoying to see something failed, but you don't know exactly what. -- Robert Blayzor, BOFH INOC, LLC rblayzor@inoc.net (A)bort, (R)etry, (P)anic?
I have had to disable oplocks also in 2.2.2 after seeing a lot of errors in the logs and having smbds which become unkillable. I thought I had solved that latter one when I got rid of the oplocks and recompiled everything (2.2.20 kernel, samba 2.2.2, ACLs) but today we had another instance when underload (about 40 users). This means the raid array cannot be unmounted and we have to shudown and reboot messily and then check all the disks etc. Not impressive. I am going to move on to 3.0 beta and run that in our production envrironment - what is there to lose now ? I just hope I can get it working so it is relatively stable. -----Original Message----- From: Robert Blayzor [mailto:noc@inoc.net] Sent: 07 January 2002 19:44 To: samba@lists.samba.org Subject: 2.2.2 Release incredibly broken? I don't know how things have been going for many of you out there with Samba 2.2.2, but I'm having horrible results with the stability of Samba in a production environment. Using Samba 2.2.2 from the FreeBSD-stable ports collection: 1) Op locks have to be disabled. There are a horrible amount of errors that Samba logs in the machine specific logs. So many that eventually smbd crashes with a signal 6. 2) When operating in domain controller mode, we've noticed a couple of times now that samba just stops working as a domain controller. No errors in the smbd or nmbd logs, it just fails to authenticate logins, etc. Restarting Samba seems to fix the problem. 3) Some Windows 2K pro machines that are member machines (with trust accounts) seem to like to blue screen when they start up. Hard crash. Usually unpluging the network jack and waiting for the machine to boot fully, then pluging it back in seems to be fine. (probably a Windows and samba problem) We also see many machine logs that have errors as follows: [2002/01/07 14:37:18, 0] rpc_server/srv_pipe.c:api_rpcTNP(1204) api_rpcTNP: api_netlog_rpc: NET_SAMLOGON failed. Which most times doesn't seem to effect anything, but it's rather annoying to see something failed, but you don't know exactly what. -- Robert Blayzor, BOFH INOC, LLC rblayzor@inoc.net (A)bort, (R)etry, (P)anic? -- To unsubscribe from this list go to the following URL and read the instructions: http://lists.samba.org/mailman/listinfo/samba
Thanks Jeremy. OK - I'll get it now and give it a whirl and get back to you at the end of the week with my findings. Noel -----Original Message----- From: jra@samba.org [mailto:jra@samba.org] Sent: 07 January 2002 22:12 To: Noel Kelly Cc: 'rblayzor@inoc.net'; samba@lists.samba.org Subject: Re: 2.2.2 Release incredibly broken? On Mon, Jan 07, 2002 at 10:05:22PM -0000, Noel Kelly wrote:> I have had to disable oplocks also in 2.2.2 after seeing a lot of errorsin> the logs and having smbds which become unkillable. I thought I had solved > that latter one when I got rid of the oplocks and recompiled everything > (2.2.20 kernel, samba 2.2.2, ACLs) but today we had another instance when > underload (about 40 users). This means the raid array cannot be unmounted > and we have to shudown and reboot messily and then check all the disksetc.> Not impressive. > > I am going to move on to 3.0 beta and run that in our production > envrironment - what is there to lose now ? I just hope I can get itworking> so it is relatively stable.That is a little radical. Currently the SAMBA_2_2 CVS Tree is being prepared for release as 2.2.3, and has fixes for (hopefully all) of the bugs the original poster and you are having problems with. Please test it out and give us feedback ASAP if it doesn't fix things, as we're in the process of doing Q/A before final release. Jeremy.