CIFS shares
Ste
125 Posts
March 11, 2020, 12:25 pmQuote from Ste on March 11, 2020, 12:25 pmHi, I just tested the CIFS shares and it works smoothly also with AD authentication, very good and useful (it saves me a file server 😉 ) ! I just have a couple of questions regarding this:
- I allocated 16 IPs in CIFS settings, and they are all active and distributed across the 3 nodes. Is there a good reason to have several IP per node, or it is enough one IP for each node, so in total to allocate as many IP as the number of cluster nodes ?
- Is there a way to limit the size of the shares, to avoid that they take all available cluster space ?
Thanks, Ste.
Hi, I just tested the CIFS shares and it works smoothly also with AD authentication, very good and useful (it saves me a file server 😉 ) ! I just have a couple of questions regarding this:
- I allocated 16 IPs in CIFS settings, and they are all active and distributed across the 3 nodes. Is there a good reason to have several IP per node, or it is enough one IP for each node, so in total to allocate as many IP as the number of cluster nodes ?
- Is there a way to limit the size of the shares, to avoid that they take all available cluster space ?
Thanks, Ste.
admin
2,930 Posts
March 11, 2020, 1:12 pmQuote from admin on March 11, 2020, 1:12 pmThanks for the feedback 🙂
you can have 1 ip per server if you wish, there is small advantage in putting more, so if you add additional servers the ips will automatically be rebalanced among them, instead of having to increase the range each time. Note we updated our quick start guide, it covers some basics to get started, but i think you do not need it by now.
The limit per share is a good idea, we took note of it for future releases.
Thanks for the feedback 🙂
you can have 1 ip per server if you wish, there is small advantage in putting more, so if you add additional servers the ips will automatically be rebalanced among them, instead of having to increase the range each time. Note we updated our quick start guide, it covers some basics to get started, but i think you do not need it by now.
The limit per share is a good idea, we took note of it for future releases.
Ste
125 Posts
March 12, 2020, 12:05 pmQuote from Ste on March 12, 2020, 12:05 pmThe limit per share is a good idea, we took note of it for future releases.
Good, I'm looking forward to it. 😉
If it might be useful, I discovered that changing CIFS IP addresses "after" joining an AD domain makes the CIFS shares unreachable (I think the AD dns is not updated). But simply un-joining and re-joining to AD fixes the issue.
Bye, Ste.
The limit per share is a good idea, we took note of it for future releases.
Good, I'm looking forward to it. 😉
If it might be useful, I discovered that changing CIFS IP addresses "after" joining an AD domain makes the CIFS shares unreachable (I think the AD dns is not updated). But simply un-joining and re-joining to AD fixes the issue.
Bye, Ste.
admin
2,930 Posts
March 12, 2020, 8:40 pmQuote from admin on March 12, 2020, 8:40 pmThis should not happen, can you please try to reproduce it. Add or delete a couple of ips and see if it occurs, note that if you change vlans or subnets it may have to restart but shoud also work in a minute or 2.
This should not happen, can you please try to reproduce it. Add or delete a couple of ips and see if it occurs, note that if you change vlans or subnets it may have to restart but shoud also work in a minute or 2.
Ste
125 Posts
March 16, 2020, 9:42 amQuote from Ste on March 16, 2020, 9:42 amOk , I made more tests and now it's more clear what happened. When the number of IPs is increased there's no outage at all, as the netbios name (i.e. PETASAN in my case) is still mapped on an active IP. When i reduce the number of IP, say from 16 down to 4, then two situations can happen:
- the netbios name was mapped on one of the unmodified IP, then everything still works;
- the netbios name was mapped on one of the deleted IP, then the shares are not reachable anymore, I see that the netbios name is still attached to the old IP, I tried also cleaning the arp table on the windows client, but with no effect. Un-joing and re-joining the storage cluster to the AD updates the mapping and solves the problem.
Anyway this is not a big issue, I just hope that this helps. 🙂 Bye.
(PS: no change in vlan or subnet)
Ok , I made more tests and now it's more clear what happened. When the number of IPs is increased there's no outage at all, as the netbios name (i.e. PETASAN in my case) is still mapped on an active IP. When i reduce the number of IP, say from 16 down to 4, then two situations can happen:
- the netbios name was mapped on one of the unmodified IP, then everything still works;
- the netbios name was mapped on one of the deleted IP, then the shares are not reachable anymore, I see that the netbios name is still attached to the old IP, I tried also cleaning the arp table on the windows client, but with no effect. Un-joing and re-joining the storage cluster to the AD updates the mapping and solves the problem.
Anyway this is not a big issue, I just hope that this helps. 🙂 Bye.
(PS: no change in vlan or subnet)
Last edited on March 16, 2020, 9:45 am by Ste · #5
admin
2,930 Posts
March 16, 2020, 12:33 pmQuote from admin on March 16, 2020, 12:33 pmActually this is as expected. If you had a client using this ip that got removed.
Note that it is your responsibility to give client valid ips. You can either assign the ips yourself to clients or connect via the Netbios name, to use the later you need to setup a round robin DNS, outside of PetaSAN. We dot provide DNS service for your clients, you need to have the DNS settings match the settings you enter in PetaSAN. There is a brief description of this in the quick start guide.
Actually this is as expected. If you had a client using this ip that got removed.
Note that it is your responsibility to give client valid ips. You can either assign the ips yourself to clients or connect via the Netbios name, to use the later you need to setup a round robin DNS, outside of PetaSAN. We dot provide DNS service for your clients, you need to have the DNS settings match the settings you enter in PetaSAN. There is a brief description of this in the quick start guide.
Last edited on March 16, 2020, 12:34 pm by admin · #6
CIFS shares
Ste
125 Posts
Quote from Ste on March 11, 2020, 12:25 pmHi, I just tested the CIFS shares and it works smoothly also with AD authentication, very good and useful (it saves me a file server 😉 ) ! I just have a couple of questions regarding this:
- I allocated 16 IPs in CIFS settings, and they are all active and distributed across the 3 nodes. Is there a good reason to have several IP per node, or it is enough one IP for each node, so in total to allocate as many IP as the number of cluster nodes ?
- Is there a way to limit the size of the shares, to avoid that they take all available cluster space ?
Thanks, Ste.
Hi, I just tested the CIFS shares and it works smoothly also with AD authentication, very good and useful (it saves me a file server 😉 ) ! I just have a couple of questions regarding this:
- I allocated 16 IPs in CIFS settings, and they are all active and distributed across the 3 nodes. Is there a good reason to have several IP per node, or it is enough one IP for each node, so in total to allocate as many IP as the number of cluster nodes ?
- Is there a way to limit the size of the shares, to avoid that they take all available cluster space ?
Thanks, Ste.
admin
2,930 Posts
Quote from admin on March 11, 2020, 1:12 pmThanks for the feedback 🙂
you can have 1 ip per server if you wish, there is small advantage in putting more, so if you add additional servers the ips will automatically be rebalanced among them, instead of having to increase the range each time. Note we updated our quick start guide, it covers some basics to get started, but i think you do not need it by now.
The limit per share is a good idea, we took note of it for future releases.
Thanks for the feedback 🙂
you can have 1 ip per server if you wish, there is small advantage in putting more, so if you add additional servers the ips will automatically be rebalanced among them, instead of having to increase the range each time. Note we updated our quick start guide, it covers some basics to get started, but i think you do not need it by now.
The limit per share is a good idea, we took note of it for future releases.
Ste
125 Posts
Quote from Ste on March 12, 2020, 12:05 pmThe limit per share is a good idea, we took note of it for future releases.
Good, I'm looking forward to it. 😉
If it might be useful, I discovered that changing CIFS IP addresses "after" joining an AD domain makes the CIFS shares unreachable (I think the AD dns is not updated). But simply un-joining and re-joining to AD fixes the issue.
Bye, Ste.
The limit per share is a good idea, we took note of it for future releases.
Good, I'm looking forward to it. 😉
If it might be useful, I discovered that changing CIFS IP addresses "after" joining an AD domain makes the CIFS shares unreachable (I think the AD dns is not updated). But simply un-joining and re-joining to AD fixes the issue.
Bye, Ste.
admin
2,930 Posts
Quote from admin on March 12, 2020, 8:40 pmThis should not happen, can you please try to reproduce it. Add or delete a couple of ips and see if it occurs, note that if you change vlans or subnets it may have to restart but shoud also work in a minute or 2.
This should not happen, can you please try to reproduce it. Add or delete a couple of ips and see if it occurs, note that if you change vlans or subnets it may have to restart but shoud also work in a minute or 2.
Ste
125 Posts
Quote from Ste on March 16, 2020, 9:42 amOk , I made more tests and now it's more clear what happened. When the number of IPs is increased there's no outage at all, as the netbios name (i.e. PETASAN in my case) is still mapped on an active IP. When i reduce the number of IP, say from 16 down to 4, then two situations can happen:
- the netbios name was mapped on one of the unmodified IP, then everything still works;
- the netbios name was mapped on one of the deleted IP, then the shares are not reachable anymore, I see that the netbios name is still attached to the old IP, I tried also cleaning the arp table on the windows client, but with no effect. Un-joing and re-joining the storage cluster to the AD updates the mapping and solves the problem.
Anyway this is not a big issue, I just hope that this helps. 🙂 Bye.
(PS: no change in vlan or subnet)
Ok , I made more tests and now it's more clear what happened. When the number of IPs is increased there's no outage at all, as the netbios name (i.e. PETASAN in my case) is still mapped on an active IP. When i reduce the number of IP, say from 16 down to 4, then two situations can happen:
- the netbios name was mapped on one of the unmodified IP, then everything still works;
- the netbios name was mapped on one of the deleted IP, then the shares are not reachable anymore, I see that the netbios name is still attached to the old IP, I tried also cleaning the arp table on the windows client, but with no effect. Un-joing and re-joining the storage cluster to the AD updates the mapping and solves the problem.
Anyway this is not a big issue, I just hope that this helps. 🙂 Bye.
(PS: no change in vlan or subnet)
admin
2,930 Posts
Quote from admin on March 16, 2020, 12:33 pmActually this is as expected. If you had a client using this ip that got removed.
Note that it is your responsibility to give client valid ips. You can either assign the ips yourself to clients or connect via the Netbios name, to use the later you need to setup a round robin DNS, outside of PetaSAN. We dot provide DNS service for your clients, you need to have the DNS settings match the settings you enter in PetaSAN. There is a brief description of this in the quick start guide.
Actually this is as expected. If you had a client using this ip that got removed.
Note that it is your responsibility to give client valid ips. You can either assign the ips yourself to clients or connect via the Netbios name, to use the later you need to setup a round robin DNS, outside of PetaSAN. We dot provide DNS service for your clients, you need to have the DNS settings match the settings you enter in PetaSAN. There is a brief description of this in the quick start guide.