Forums

Home / Forums

You need to log in to create posts and topics. Login · Register

i am new here command line

Pages: 1 2 3 4

admin
2,930 Posts
February 5, 2020, 12:04 pm

can you try to ping on backend networks from node 2 to 1.

double check your connections, switch configuration.

 

#21

Leobosch1
19 Posts
February 14, 2020, 1:55 pm

Thanks Admin,

i am really enjoying the usage of petaSAN, i found out that one of my node in my cluster automatically log out other nodes (4 nodes in total)

what is the function of fencing in maintenance?

Thanks

#22

admin
2,930 Posts
February 14, 2020, 8:41 pm

Fencing always a node to kill/shutdown another node if the later is not responding to cluster heartbeats. before faillover and taking over resources such as ip and disk access, this is to make sure the node is dead.

You can disable this from the maintenance tab (not recommended)

if you find this happens frequently, it is better to know why the heartbeats are not replied too in time, this could be hardware or network problems or hardware is not powerful enough for the load.

Last edited on February 14, 2020, 8:42 pm by admin · #23

Leobosch1
19 Posts
February 21, 2020, 11:48 am

how can i change the health of my cluster when it display (too many PGs per OSD (576 > max 300)?

Thanks

Last edited on February 21, 2020, 11:48 am by Leobosch1 · #24

admin
2,930 Posts
February 21, 2020, 2:09 pm

You have too many PG in you pools than OSDs. Maybe you added more pools yourself or chose an incorrect OSD count when creating the cluster default pool.

to  fix: either add more OSDs ( double the current size ), or if you do not need current pools (and data) , delete them and create new pools with less PGs.

Last edited on February 21, 2020, 2:09 pm by admin · #25

Leobosch1
19 Posts
April 6, 2020, 11:21 am

Hello

i am having a problem i am trying to update from 2.3.1 to 2.5.1 but i have 4 node i followed the online upgrade instructions per node but only one of the node was upgrade i try to restart the the other nodes that was not upgraded but i was getting  2.3.1 only one of the Node which was the first one was upgraded so out of my 4 nodes it was only 1 that was upgraded. i try it again on the remaining three node from the beginning but i was then having a problem like this "Incorrect Ceph version found." what must i do now? i will be grateful if you can help me.

Thanks

 

 

 

#26

admin
2,930 Posts
April 6, 2020, 2:44 pm

was the cluster installed new in 2.3.1 or was it upgraded from earlier versions.

in one of the nodes with issue, what is the output of

ceph status

dpkg -l | grep ceph

dpkg -l | grep petasan

#27

Leobosch1
19 Posts
April 6, 2020, 6:15 pm

Hello

i have 4 nodes with 2.3.1 installed originally  but i tried to upgrade to 2.5.1 but i did the node online upgrade with all the nodes but only 1 node was successful.

so the remaining 3 nodes  was not successful is still 2.3.1 i tried repeating the upgrade only for this 3 nodes so that i can make these 3 nodes to be like the 1 node but it is showing me  "Incorrect Ceph version found."

the three of the nodes with issue, what is the output of

ceph status

ok and healthy

dpkg -l | grep ceph

14.2.7-1petasan

dpkg -l | grep petasan

14.2.7-1petasan

what must i do now?

Thanks

#28

admin
2,930 Posts
April 6, 2020, 7:05 pm

did you run

apt install petasan

if so, what does it say ?

what is output of

dpkg -s petasan | grep Version

Last edited on April 6, 2020, 7:09 pm by admin · #29

Leobosch1
19 Posts
April 6, 2020, 7:11 pm

yes i did the following

apt update

export DEBIAN_FRONTEND=noninteractive

apt upgrade

apt install petasan

Thanks

 

#30
Pages: 1 2 3 4