hunterzzz
1 Copper

How can i restart the SMB services on a specific node of the cluster

Jump to solution

we are experiencing some issues with a number of nodes in our cluster where we are unable to connect to the administrative share (i.e. ifs$). I tested connecting to this share using the individual IP addresses of each node and find that i am unable to connect to 5 of them, and there are no SMB connections reported via the OneFS GUI. So, i choose one of the nodes giving issues, and i was wondering just how i could restart the SMB services on it to see if that would resolve the issue, I would rather not reboot the node as there are a few NFS connections present. We are currently running OneFS v7.0.1.3

I have attached a screen shot of what i am seeing.

Tags (3)
0 Kudos
1 Solution

Accepted Solutions
Highlighted
soetingr
1 Nickel

Re: How can i restart the SMB services on a specific node of the cluster

Jump to solution

Hi,

On the node you are suspecting login with ssh and type:

node-1# ps -aux|grep lw

You should see 6 processes like this:

root    2758 18.0  0.1 174888 28844  ??  S    12Feb13 1006:49.92 lw-container lsass (lsass)

root    2325  0.0  0.0 93296  7408  ??  I    12Feb13   0:00.19 /usr/likewise/sbin/lwsmd --syslog --start-as-daemon

root    2682  0.0  0.1 110392 12720  ??  I    12Feb13   0:21.21 lw-container lwreg (lwreg)

root    2700  0.0  0.2 142036 52256  ??  I    12Feb13   0:26.68 lw-container netlogon (netlogon)

root    2747  0.0  0.1 120332 13684  ??  I    12Feb13   0:04.34 lw-container lwio (lwio)

root    3039  0.0  0.0 103868 11916  ??  I    12Feb13   0:05.59 lw-container srvsvc (srvsvc)

root   55912  0.0  0.0  7960  1288  p0  S+   10:58AM   0:00.00 grep lw

If not or you don't see all of them, use killall to kill the local process on that node.

node-1# killall srvsvc

It takes some time, but a background process should restart the process.

On a second ssh session run:

isi statistics system --nodes --top

This will show you that after a while smb sessions get access to this node again.


We had the a similar problem (bug record #99264). We are planning a upgrade to 7.0.2.1.


Regards,

Robert

0 Kudos
3 Replies
christopher_ime
4 Beryllium

Re: How can i restart the SMB services on a specific node of the cluster

Jump to solution

hunterzzz,

Firstly, please consider moving this question to the proper forum for maximum visibility.  Questions written to the users' own "Discussions" space don't get the same amount of attention and questions can go unanswered for a long time.  You can do so by selecting "Move" under ACTIONS along the right.  Then search for and select: "Isilon Support Forum" which would be the most relevant for this question

I wonder if maybe you are experiencing the issue as documented in the following KB article.  You are running an affected version of OneFS; well... at least when you had posted this question.

emc321867: "Locking issue can cause SMB connections to become non-responsive"

As noted, if you can't upgrade to the latest OneFS (v7.0.1.6 as of this reply) or at a minimum install the patch, a possible work-around is to restart the lwiod daemon on the affected node.

0 Kudos
christopher_ime
4 Beryllium

Re: How can i restart the SMB services on a specific node of the cluster

Jump to solution

hunterzzz,

I simply wanted to follow-up with you to make sure you get the best possible experience from the forums and the community of dedicated customers, partners, and EMC employees that are eager to assist.

1) As suggested, please consider relocating this post so that it gets maximum visibility and therefore assures that you get the most (quality and quantity) feedback by the community.  The most relevant support community for this question would be: "Isilon Support Forum".  Here is a recent post providing detailed instructions on how to accomplish this:

EMC Community Network Frequently Asked Questions

2) Once relocated, if/when answered please also mark as such, so that it may help others that seek the same advice

0 Kudos
Highlighted
soetingr
1 Nickel

Re: How can i restart the SMB services on a specific node of the cluster

Jump to solution

Hi,

On the node you are suspecting login with ssh and type:

node-1# ps -aux|grep lw

You should see 6 processes like this:

root    2758 18.0  0.1 174888 28844  ??  S    12Feb13 1006:49.92 lw-container lsass (lsass)

root    2325  0.0  0.0 93296  7408  ??  I    12Feb13   0:00.19 /usr/likewise/sbin/lwsmd --syslog --start-as-daemon

root    2682  0.0  0.1 110392 12720  ??  I    12Feb13   0:21.21 lw-container lwreg (lwreg)

root    2700  0.0  0.2 142036 52256  ??  I    12Feb13   0:26.68 lw-container netlogon (netlogon)

root    2747  0.0  0.1 120332 13684  ??  I    12Feb13   0:04.34 lw-container lwio (lwio)

root    3039  0.0  0.0 103868 11916  ??  I    12Feb13   0:05.59 lw-container srvsvc (srvsvc)

root   55912  0.0  0.0  7960  1288  p0  S+   10:58AM   0:00.00 grep lw

If not or you don't see all of them, use killall to kill the local process on that node.

node-1# killall srvsvc

It takes some time, but a background process should restart the process.

On a second ssh session run:

isi statistics system --nodes --top

This will show you that after a while smb sessions get access to this node again.


We had the a similar problem (bug record #99264). We are planning a upgrade to 7.0.2.1.


Regards,

Robert

0 Kudos