Unsolved
1 Rookie
•
17 Posts
0
1168
February 24th, 2021 04:00
nfs clients hung during OneFS Upgrade
During a rolling reboot a couple of NFS clients reported issue as “nfs mount points hung/df –k not responding” and rebooted the client nodes to fix the issue.
Clients were running RHEL 7.6 and NFS v3
Mount paramters
rw,relatime,vers=3,rsize=131072,wsize=524288,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=x.x.x.x,mountvers=3,mountport=300,mountproto=udp,local_lock=none,addr=x.x.x.x
The network pool is set to dynamic allocation
As I understand it the nfs clients should not see disruption during a rolling reboot?
The users are keen to understand why this occurred and to prevent a recurrance during a planned OneFS upgrade
0 events found
No Events found!
DELL-Josh Cr
Moderator
•
9.6K Posts
0
February 24th, 2021 11:00
Hi,
What version were you upgrading from and to? It sounds like they were not disconnecting and reconnecting properly. https://dell.to/3pMAH4x
paf23
1 Rookie
•
17 Posts
0
February 25th, 2021 02:00
Hi - thanks for the reply
The upgrade being performed was node and disk firmware. Currently running OneFS 8.1.2 and planning an upgrade to 8.2.2. The customer is concerned that they may see more disruption during the planned OneFS upgrade
We performed a rolling reboot before the upgrade as the cluster had not been rebooted for a long time.
We then did the node firmware upgrade and then the disk firmware.
DELL-Sam L
Moderator
•
7.9K Posts
0
February 25th, 2021 10:00
Hello paf23,
Here is a link to a couple of additional KB that maybe of assistance. https://dell.to/3qWg0V9
https://dell.to/2MqchA9
Phil.Lam
3 Apprentice
•
637 Posts
0
February 25th, 2021 15:00
@paf23,
Please open an SR to investigate further. Some possible KBs.
531354 : {ISILON} Intermittent disconnection to Isilon nodes' dynamic IPs during rolling reboot or rolling upgrade. https://support.emc.com/kb/531354
490729 : Clients disconnect after IP balance when Flooding Mode turned off on Cisco Nexus switches https://support.emc.com/kb/490729
paf23
1 Rookie
•
17 Posts
0
March 1st, 2021 02:00
Thanks everybody for the information. I am checking with my customer to see if the KB articles are relevant to their environment.
paf23
1 Rookie
•
17 Posts
0
March 3rd, 2021 03:00
The customer has stated that they are not using ACI - both of the supplied KB articles seem to reference ACI on the Cisco side.
I will look at raising an SR.
cadencep45
3 Apprentice
•
318 Posts
0
March 3rd, 2021 07:00
ensure they put mount as defined in /etc/fstab so a static rather than pool reference can be confirmed. I completed an isilon upgrade to 9.x with similar clients with no reboots encountered.
cadencep45
3 Apprentice
•
318 Posts
0
March 3rd, 2021 07:00
worst case, the nfs mount should hang while the node is being rebooted, a client reboot is not expected behavior