New CentOS 7 file server, samba version 4.10.4-11 with 2,000 shares on multiple Gluster mounts. Roughly 200-300 concurrent users. We recently migrated this working setup from a CentOS 5 server (Samba 3) using LDAP. A few weeks later we started noticing sporadic slowdowns on our Windows 10 PC clients. Directory listings that normally take a few seconds can take many minutes if they finish at all. Nothing jumps out in the Samba logs. A local linux find on the affected shares does not vary much, about a minute. I've run smbclient 'recurse;ls' on the share(s) both locally and from another Linux host. Local and remote search times (good and bad) agree closely: A few minutes is normal but can take an hour or more when we're having problems. Gluster or network is not suspect. Sometimes cycling smb fixes it for a while but not today. This happens mostly during business hours but the server is always 95% idle. Any ideas or what to look for in the logs would be appreciated. smb.conf: [global] kerberos method = system keytab log level = 3 max log size = 5000 log file = /var/log/samba/log.%h.%m template homedir = /home/%U@%D template shell = /bin/bash security = ads realm = MYGROUP.COM idmap config MYGROUP : range = 1034-2999999 idmap config MYGROUP : backend = ad idmap config MYGROUP : schema_mode = rfc2307 idmap config MYGROUP : unix_primary_group = yes idmap config MYGROUP : unix_nss_info = yes idmap config * : range = 10000-15000 idmap config * : backend = tdb winbind use default domain = yes winbind refresh tickets = yes winbind offline logon = yes winbind enum groups = no winbind enum users = no workgroup = MYGROUP realm = MYGROUP.COM include = /usr/local/custom.conf Thanks, Jim Brand This email and any attachments may contain information that is confidential and/or privileged for the sole use of the intended recipient. Any use, review, disclosure, copying, distribution or reliance by others, and any forwarding of this email or its contents, without the express permission of the sender is strictly prohibited by law. If you are not the intended recipient, please contact the sender immediately, delete the e-mail and destroy all copies.
On Wed, 2021-06-30 at 19:45 +0000, Jim Brand via samba wrote:> New CentOS 7 file server, samba version 4.10.4-11 with 2,000 shares > on multiple Gluster mounts. Roughly 200-300 concurrent users. We > recently migrated this working > setup from a CentOS 5 server (Samba 3) using LDAP. A few weeks > later we started noticing sporadic slowdowns on our Windows 10 PC > clients. > Directory listings that normally take a few seconds can take many > minutes if they finish at all. > > Nothing jumps out in the Samba logs. A local linux find on the > affected shares does not vary much, about a minute. I've run > smbclient 'recurse;ls' on the share(s) both locally and > from another Linux host. Local and remote search times (good and > bad) agree closely: A few minutes is normal but can take an hour or > more when we're having problems. Gluster or network is not > suspect. Sometimes cycling smb fixes it for a while but not today. > > This happens mostly during business hours but the server is always > 95% idle. Any ideas or what to look for in the logs would be > appreciated. > > smb.conf: > > [global] > idmap config MYGROUP : range = 1034-2999999 > idmap config * : range = 10000-15000Your ranges overlap, this isn't allowed. How did you migrate from the old server ? I suppose that I should also mention that from a Samba point of view 4.10.x is EOL. Rowland
On Wed, Jun 30, 2021 at 07:45:42PM +0000, Jim Brand via samba wrote:>New CentOS 7 file server, samba version 4.10.4-11 with 2,000 shares on multiple Gluster mounts. Roughly 200-300 concurrent users. We recently migrated this working >setup from a CentOS 5 server (Samba 3) using LDAP. A few weeks later we started noticing sporadic slowdowns on our Windows 10 PC clients. >Directory listings that normally take a few seconds can take many minutes if they finish at all. > >Nothing jumps out in the Samba logs. A local linux find on the affected shares does not vary much, about a minute. I've run smbclient 'recurse;ls' on the share(s) both locally and >from another Linux host. Local and remote search times (good and bad) agree closely: A few minutes is normal but can take an hour or more when we're having problems. Gluster or network is not suspect. Sometimes cycling smb fixes it for a while but not today. > >This happens mostly during business hours but the server is always 95% idle. Any ideas or what to look for in the logs would be appreciated. > >smb.conf: > >[global] >kerberos method = system keytab >log level = 3 >max log size = 5000 >log file = /var/log/samba/log.%h.%m >template homedir = /home/%U@%D >template shell = /bin/bash >security = ads >realm = MYGROUP.COM >idmap config MYGROUP : range = 1034-2999999 >idmap config MYGROUP : backend = ad >idmap config MYGROUP : schema_mode = rfc2307 >idmap config MYGROUP : unix_primary_group = yes >idmap config MYGROUP : unix_nss_info = yes >idmap config * : range = 10000-15000 >idmap config * : backend = tdb >winbind use default domain = yes >winbind refresh tickets = yes >winbind offline logon = yes >winbind enum groups = no >winbind enum users = no >workgroup = MYGROUP >realm = MYGROUP.COM > >include = /usr/local/custom.conf > >Thanks, >Jim BrandCan you reproduce this from the Windows clients using a command line listing, or only from Windows explorer ? I think explorer gets more meta-data than a raw cmd.exe "DIR" will do. That might give a hint on where to look.