[Samba] IO report error when use samba in HA

[Stefan Kania]

oh yes, there are some more parameters you need in Gluster. Here are a 
list of all parameters:
--------------------
performance.read-ahead=on
performance.write-behind-window-size=4MB
performance.cache-max-file-size=10
cluster.force-migration=on
cluster.entry-self-heal=on
cluster.metadata-self-heal=on
cluster.data-self-heal=on
cluster.self-heal-daemon=enable
features.cache-invalidation=on
features.cache-invalidation-timeout=600
performance.cache-samba-metadata=on
performance.stat-prefetch=on
performance.md-cache-timeout=600
network.inode-lru-limit=200000
performance.nl-cache=on
performance.nl-cache-timeout=600
performance.readdir-ahead=on
performance.parallel-readdir=on
client.event-threads=4
server.event-threads=4
performance.cache-invalidation=on
performance.write-behind=off
network.ping-timeout=10
performance.cache-size=512MB
user.cifs=enable
user.smb=enable
--------------------
store these parametes in a file for example "mysamba" then copy the file 
to /var/lib/glusterd/groups then do a:
gluster v set vol-name group mysamba
this will set all the parameters from the list. As you can see there are 
some caching parameters if you have a large amount of RAM you can change 
these values, but I would start with theses settings ;-)


Am 03.11.22 um 13:08 schrieb jie wang:
> Thanks the reply, I set network.ping.timeout to 10, but it does not
> work. Following are my volume's parameters, did I need to change other
> parameter ?
> 
> gluster> volume info
> Volume Name: tank
> Type: Distribute
> Volume ID: 6ab518a2-08c1-4456-af1b-31a47926d246
> Status: Started
> Snapshot Count: 0
> Number of Bricks: 3
> Transport-type: tcp
> Bricks:
> Brick1: 11.10.10.10:/data01/gluster
> Brick2: 11.10.10.11:/data01/gluster
> Brick3: 11.10.10.12:/data01/gluster
> Options Reconfigured:
> network.ping-timeout: 10
> performance.write-behind: off
> transport.address-family: inet
> performance.readdir-ahead: on
> nfs.disable: on
> 
> On Thu, Nov 3, 2022 at 7:29 PM Stefan Kania via samba
> <samba at lists.samba.org> wrote:
>>
>> Did you configure the right parameters for your GlusterFS to fully
>> support Samba? Did you change the network.ping.timeout from Gluster? The
>> default value is 42 seconds and it's much to long for Windows-Clients
>> you should have no more then 20 seconds set if you use Gluster with Windows.
>>
>> Am 03.11.22 um 02:17 schrieb jie wang via samba:
>>> Hi all,
>>>     I use two samba server1 + glusterfs + ctdb to build a high availability
>>> environment, and use windows client to visit the samba server. First,
>>> windows client connect to samba server1,  and I execute a copy command
>>> in client which cost a long time. Then I kill samba server1, and
>>> windows client reconnect to samba server2, but the copy command break
>>> and report error. But the strange thing is, if I use linux client, the
>>> copy command will not break, in the period of reconnect.
>>>     So how can I make the windows client not break the copy command, in
>>> the period of reconnect ? Thanks a lot.
>>>
>>
>>
>>
>>
>> --
>> To unsubscribe from this list go to the following URL and read the
>> instructions:  https://lists.samba.org/mailman/options/samba

-- 
Stefan Kania
Landweg 13
25693 St. Michaelisdonn


Signieren jeder E-Mail hilft Spam zu reduzieren und schützt Ihre 
Privatsphäre. Ein kostenfreies Zertifikat erhalten Sie unter 
https://www.dgn.de/dgncert/index.html