# SLI IAD GlusterFS nodes retype m5 -> m6i
```
0. Preparation
[X] 0.1. Instances summary sheet https://docs.google.com/spreadsheets/d/1AXKJ5AkBpIq9pGRKOdxF0LDwPOWOmwp_luWqudt1cLI/edit#gid=0
[X] 0.2. Ensure accesses to all instances
[X] 0.3. Perform test run on a copy of any clusters.
Done, using IAD GFS instances' backup AMIs.
[X]
```
```
1. Pre-actions check Cluster 1 (IAD)
Common approach is to check cluster for any errors before performing any modifications.
If issues are found (except for easy tasks like mount/remount volumes), escalate and postpone executions.
[X] 1.1. Check nagios for issues related to gfs*.iad* hosts
[X] 1.2. Ensure instance i-0af638f14f94e3ce0 (gfs1-1.iad-aws.prod.sli.io) is ready for retype.
[X] 1.2.0. Connect to the instance:
ssh 10.79.114.41
[X] 1.2.1. Check mount:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-1.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-1.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
[X] 1.2.2. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 1.2.1.
[X] 1.2.3. Check node VIP status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL OK - IN VIP: write_file,success; read_file,success; alive;
If output of the command is as shown above, proceed to the next step.
[root@gfs1-1.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL CRITICAL - NOT IN VIP: write_file,failed; check_failed; remove,already_removed;
If output of the command shows above failure, try following:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ vip-control insert
Sat Jul 31 00:38:39 NZST 2021,vip-control,insert,successful
If above command succeeded as shown, continue to the next step.
If fails, stop execution and escalate to BU SaaS.
[X] 1.2.4. Check peers status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
If there are issues with nodes above (not Connected), stop execution and escalate to BU SaaS.
[X] 1.2.5. Check volumes status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are not online, stop execution and escalate to BU SaaS.
If there are active volume tasks in progress, wait for completion before continue with modifications.
[up to 30min, then escalate]
```
```
[X] 1.3. Ensure instance i-08738fdcf35071be8 (gfs1-2.iad-aws.prod.sli.io) is ready for retype.
[X] 1.3.0. Connect to the instance:
ssh 10.79.114.45
[X] 1.3.1. Check mount:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-2.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-2.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
[X] 1.3.2. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 1.3.1.
[X] 1.3.3. Check node VIP status:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL OK - IN VIP: write_file,success; read_file,success; alive;
If output of the command is as shown above, proceed to the next step.
[root@gfs1-2.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL CRITICAL - NOT IN VIP: write_file,failed; check_failed; remove,already_removed;
If output of the command shows above failure, try following:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ vip-control insert
Sat Jul 31 00:38:39 NZST 2021,vip-control,insert,successful
If above command succeeded as shown, continue to the next step.
If fails, stop execution and escalate to BU SaaS.
[X] 1.3.4. Check peers status:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
If there are issues with nodes above (not Connected), stop execution and escalate to BU SaaS.
[X] 1.3.5. Check volumes status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are not online, stop execution and escalate to BU SaaS.
If there are active volume tasks in progress, wait for completion before continue with modifications.
[up to 30min, then escalate]
```
```
[X] 1.4. Ensure instance i-0a0c2ad3bba430706 (gfs1-3.iad-aws.prod.sli.io) is ready for retype.
[X] 1.4.0. Connect to the instance:
ssh 10.79.114.199
[X] 1.4.1. Check mount:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-3.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-3.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
[X] 1.4.2. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 1.4.1.
[X] 1.4.3. Check node VIP status:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL OK - IN VIP: write_file,success; read_file,success; alive;
If output of the command is as shown above, proceed to the next step.
[root@gfs1-3.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL CRITICAL - NOT IN VIP: write_file,failed; check_failed; remove,already_removed;
If output of the command shows above failure, try following:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ vip-control insert
Sat Jul 31 00:38:39 NZST 2021,vip-control,insert,successful
If above command succeeded as shown, continue to the next step.
If fails, stop execution and escalate to BU SaaS.
[X] 1.4.4. Check peers status:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
If there are issues with nodes above (not Connected), stop execution and escalate to BU SaaS.
[X] 1.4.5. Check volumes status:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are not online, stop execution and escalate to BU SaaS.
If there are active volume tasks in progress, wait for completion before continue with modifications.
[up to 30min, then escalate]
```
```
[X] 1.5. Ensure instance i-001707ebdd3539c48 (gfs1-4.iad-aws.prod.sli.io) is ready for retype.
[X] 1.5.0. Connect to the instance:
ssh 10.79.114.67
[X] 1.5.1. Check mount:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-4.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-4.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
[X] 1.5.2. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 1.5.1.
[X] 1.5.3. Check node VIP status:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL OK - IN VIP: write_file,success; read_file,success; alive;
If output of the command is as shown above, proceed to the next step.
[root@gfs1-4.iad-aws.prod.sli.io ~]$ vip-control status
VIP-CONTROL CRITICAL - NOT IN VIP: write_file,failed; check_failed; remove,already_removed;
If output of the command shows above failure, try following:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ vip-control insert
Sat Jul 31 00:38:39 NZST 2021,vip-control,insert,successful
If above command succeeded as shown, continue to the next step.
If fails, stop execution and escalate to BU SaaS.
[X] 1.5.4. Check peers status:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
If there are issues with nodes above (not Connected), stop execution and escalate to BU SaaS.
[X] 1.5.5. Check volumes status:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 77317
Self-heal Daemon on gfs1-1.iad.prod.sli.io N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are not online, stop execution and escalate to BU SaaS.
If there are active volume tasks in progress, wait for completion before continue with modifications.
[up to 30min, then escalate]
```
```
2. Processing Cluser 1 (IAD)
[X] 2.0. Remove Cluster 1 (IAD) out of rotation:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do ssh ${host} 'runuser -l gbrain bash -c "webcontrol vip remove"'; done
(allow a minute before performing next step)
[X] 2.1. Ensure Cluster 1 (IAD) is not serving to the clients
2.1.1. Check vip status on searches:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do ssh ${host} 'runuser -l gbrain bash -c "webcontrol vip status"'; done
2.1.2. Request healthcheck document from the search hosts:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo -n "${host}: "; curl -A "Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:59.0) Gecko/20100101 Firefox/59.0" http://${host}/shared/vip-status.txt; done
expecting all hosts to return down
2.2. Perform retype of all cluster instances
[X] 2.2.1. Switch off CRON on all GFS instances
[local]% for host in 10.79.114.41 10.79.114.45 10.79.114.199 10.79.114.67; do ssh ${host} "systemctl disable crond; systemctl stop crond; pkill -f cron -9; pkill -f CRON -9; ps ax|grep -i cron"; done
[X] 2.2.2. Remove all GFS instances out of serving requests
[local]% for host in 10.79.114.41 10.79.114.45 10.79.114.199 10.79.114.67; do ssh ${host} "vip-control remove"; done
2.2.3. Proceed with instance retype
[X] 2.2.3.1. gfs1-1.iad-aws.prod.sli.io
[local]% ./automation/bin/run.sh arn:aws:ssm:us-east-1:862508395508:document/co-wu-resize-instance "InstanceId=i-0af638f14f94e3ce0,NewInstanceType=m6i.large,SkipComputeOptimizerChecks=true" 610092376560 us-east-1
[X] 2.2.3.1. gfs1-2.iad-aws.prod.sli.io
[local]% ./automation/bin/run.sh arn:aws:ssm:us-east-1:862508395508:document/co-wu-resize-instance "InstanceId=i-08738fdcf35071be8,NewInstanceType=m6i.large,SkipComputeOptimizerChecks=true" 610092376560 us-east-1
[X] 2.2.3.1. gfs1-3.iad-aws.prod.sli.io
[local]% ./automation/bin/run.sh arn:aws:ssm:us-east-1:862508395508:document/co-wu-resize-instance "InstanceId=i-0a0c2ad3bba430706,NewInstanceType=m6i.large,SkipComputeOptimizerChecks=true" 610092376560 us-east-1
[X] 2.2.3.1. gfs1-4.iad-aws.prod.sli.io
[local]% ./automation/bin/run.sh arn:aws:ssm:us-east-1:862508395508:document/co-wu-resize-instance "InstanceId=i-001707ebdd3539c48,NewInstanceType=m6i.large,SkipComputeOptimizerChecks=true" 610092376560 us-east-1
[X] 2.2.4. Return all GFS instances to serve requests
[local]% for host in 10.79.114.41 10.79.114.45 10.79.114.199 10.79.114.67; do echo -n "${host}: "; ssh ${host} "vip-control insert"; done
all hosts expected to return successful result
Sat Jul 31 00:38:39 NZST 2021,vip-control,insert,successful
if any host will return failure, investigate individual host
[X] 2.3. Perform GFS healthcheck for instance i-0af638f14f94e3ce0 (gfs1-1.iad-aws.prod.sli.io).
[X] 2.3.0. Connect to the instance:
ssh 10.79.114.41
[X] 2.3.1. Check gluster daemon is running:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ systemctl status glusterd
If not started, (re)start glusterd
[root@gfs1-1.iad-aws.prod.sli.io ~]$ systemctl restart glusterd
If still fail to start all daemons (glusterd, glusterfsd, glusterfs), investigate logs in /var/log/glusterfs/
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.3.2. Check mount:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-1.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-1.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If mounts are missing, try to mount volumes:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
If mount fails, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.3.3. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-1.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-1.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 2.3.2.
If fails for the second time, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.3.4. Check volumes status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are offline, perform following on failed node:
[local]% ssh <IP address of the node>
[root@gfs1-<node ID>.iad-aws.prod.sli.io ~]$ gluster volume start <volume name> force
If above command's output will state that there are active volume tasks, continue, but doublecheck in 1h and notify BU SaaS.
[X] 2.3.5. Check peers' status:
[root@gfs1-1.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
If there are issues in above output, something unexpected happened. As if all bricks are online on previous step, but some of peers are disconnected, this is totally wrong. So escalate to BU SaaS and continue.
[X] 2.3.6. Switch on CRON
[root@gfs1-1.iad-aws.prod.sli.io ~]$ systemctl enable crond
[root@gfs1-1.iad-aws.prod.sli.io ~]$ systemctl start crond
[X] 2.4. Perform GFS healthcheck for instance i-08738fdcf35071be8 (gfs1-2.iad-aws.prod.sli.io).
[X] 2.4.0. Connect to the instance:
ssh 10.79.114.45
[X] 2.4.1. Check gluster daemon is running:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ systemctl status glusterd
If not started, (re)start glusterd
[root@gfs1-2.iad-aws.prod.sli.io ~]$ systemctl restart glusterd
If still fail to start all daemons (glusterd, glusterfsd, glusterfs), investigate logs in /var/log/glusterfs/
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.4.2. Check mount:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-2.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-2.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If mounts are missing, try to mount volumes:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
If mount fails, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.4.3. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-2.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 2.4.2.
If fails for the second time, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.4.4. Check volumes status:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are offline, perform following on failed node:
[local]% ssh <IP address of the node>
[root@gfs1-<node ID>.iad-aws.prod.sli.io ~]$ gluster volume start <volume name> force
If above command's output will state that there are active volume tasks, continue, but doublecheck in 1h and notify BU SaaS.
[X] 2.4.5. Check peers' status:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
If there are issues in above output, something unexpected happened. As if all bricks are online on previous step, but some of peers are disconnected, this is totally wrong. So escalate to BU SaaS and continue.
[X] 2.4.6. Switch on CRON
[root@gfs1-2.iad-aws.prod.sli.io ~]$ systemctl enable crond
[root@gfs1-2.iad-aws.prod.sli.io ~]$ systemctl start crond
[X] 2.5. Perform GFS healthcheck for instance i-0a0c2ad3bba430706 (gfs1-3.iad-aws.prod.sli.io).
[X] 2.5.0. Connect to the instance:
ssh 10.79.114.199
[X] 2.5.1. Check gluster daemon is running:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ systemctl status glusterd
If not started, (re)start glusterd
[root@gfs1-3.iad-aws.prod.sli.io ~]$ systemctl restart glusterd
If still fail to start all daemons (glusterd, glusterfsd, glusterfs), investigate logs in /var/log/glusterfs/
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.5.2. Check mount:
[root@gfs1-2.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-3.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-3.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If mounts are missing, try to mount volumes:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
If mount fails, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.5.3. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-3.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-3.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 2.5.2.
If fails for the second time, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.5.4. Check volumes status:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are offline, perform following on failed node:
[local]% ssh <IP address of the node>
[root@gfs1-<node ID>.iad-aws.prod.sli.io ~]$ gluster volume start <volume name> force
If above command's output will state that there are active volume tasks, continue, but doublecheck in 1h and notify BU SaaS.
[X] 2.5.5. Check peers' status:
[root@gfs1-3.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-4.iad-aws.prod.sli.io
Uuid: 0183b376-1aa0-4d98-945e-6197a8876f27
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
If there are issues in above output, something unexpected happened. As if all bricks are online on previous step, but some of peers are disconnected, this is totally wrong. So escalate to BU SaaS and continue.
[X] 2.5.6. Switch on CRON
[root@gfs1-3.iad-aws.prod.sli.io ~]$ systemctl enable crond
[root@gfs1-3.iad-aws.prod.sli.io ~]$ systemctl start crond
[X] 2.6. Perform GFS healthcheck for instance i-001707ebdd3539c48 (gfs1-4.iad-aws.prod.sli.io).
[X] 2.6.0. Connect to the instance:
ssh 10.79.114.67
[X] 2.6.1. Check gluster daemon is running:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ systemctl status glusterd
If not started, (re)start glusterd
[root@gfs1-4.iad-aws.prod.sli.io ~]$ systemctl restart glusterd
If still fail to start all daemons (glusterd, glusterfsd, glusterfs), investigate logs in /var/log/glusterfs/
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.6.2. Check mount:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount|grep gfs1
/dev/mapper/vg01-gfs1.brick1 on /gluster/gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
/gluster/sec-gfs1/sec-gfs1.brick1 on /gluster/sec-gfs1/brick1 type xfs (rw,relatime,attr2,inode64,noquota)
gfs1-4.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1-4.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If mounts are missing, try to mount volumes:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
If mount fails, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.6.3. Check mount is healthy:
Following commands should produce no output:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
If nothing printed out, proceed to the next step.
Following is an example of mounted volume, which has issues:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1: Transport endpoint is not connected
or
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
[root@gfs1-4.iad-aws.prod.sli.io ~]$ ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Try easy solution first:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ umount /mnt/gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ umount /mnt/sec-gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/gfs1
[root@gfs1-4.iad-aws.prod.sli.io ~]$ mount /mnt/sec-gfs1
After issuing above commands, retry whole step starting with 2.6.2.
If fails for the second time, investigate logs in /var/log/glusterfs/, /var/log/messages, dmesg
[If no solution found in 10 minutes, escalate to BU SaaS]
[X] 2.6.4. Check volumes status:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ gluster volume status
Status of volume: gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3510
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31340
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 3316
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/g
fs1/brick1 49152 0 Y 31413
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Task Status of Volume gfs1
------------------------------------------------------------------------------
There are no active volume tasks
Status of volume: sec-gfs1
Gluster process TCP Port RDMA Port Online Pid
------------------------------------------------------------------------------
Brick gfs1-1.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3521
Brick gfs1-2.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31351
Brick gfs1-3.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 3339
Brick gfs1-4.iad-aws.prod.sli.io:/gluster/s
ec-gfs1/brick1 49153 0 Y 31424
Self-heal Daemon on localhost N/A N/A Y 3665
Self-heal Daemon on gfs1-3.iad-aws.prod.sli
.io N/A N/A Y 20255
Self-heal Daemon on gfs1-4.iad-aws.prod.sli
.io N/A N/A Y 66457
Self-heal Daemon on gfs1-2.iad-aws.prod.sli
.io N/A N/A Y 77317
Task Status of Volume sec-gfs1
------------------------------------------------------------------------------
There are no active volume tasks
If any of bricks above are offline, perform following on failed node:
[local]% ssh <IP address of the node>
[root@gfs1-<node ID>.iad-aws.prod.sli.io ~]$ gluster volume start <volume name> force
If above command's output will state that there are active volume tasks, continue, but doublecheck in 1h and notify BU SaaS.
[X] 2.6.5. Check peers' status:
[root@gfs1-4.iad-aws.prod.sli.io ~]$ gluster peer status
Number of Peers: 3
Hostname: gfs1-3.iad-aws.prod.sli.io
Uuid: 23884f97-8149-4dd9-b4d3-1faccee03d1a
State: Peer in Cluster (Connected)
Hostname: gfs1-2.iad-aws.prod.sli.io
Uuid: 95e9d91b-fe51-4e0d-86cd-00097890f113
State: Peer in Cluster (Connected)
Hostname: gfs1-1.iad.prod.sli.io
Uuid: c9381ca4-99ac-4e69-8823-b9b8e208f873
State: Peer in Cluster (Connected)
If there are issues in above output, something unexpected happened. As if all bricks are online on previous step, but some of peers are disconnected, this is totally wrong. So escalate to BU SaaS and continue.
[X] 2.6.6. Switch on CRON
[root@gfs1-4.iad-aws.prod.sli.io ~]$ systemctl enable crond
[root@gfs1-4.iad-aws.prod.sli.io ~]$ systemctl start crond
```
```
2.7. Checking remain cluster components' health
2.7.1. Search servers
[X] 2.7.1.1. Ensure mounts
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo ${host}; ssh ${host} "mount | grep gfs1"; done
all hosts should report both gfs1 and sec-gfs1 are mounted
systemd-1 on /mnt/sec-gfs1 type autofs (rw,relatime,fd=30,pgrp=1,timeout=0,minproto=5,maxproto=5,direct,pipe_ino=10485)
systemd-1 on /mnt/gfs1 type autofs (rw,relatime,fd=32,pgrp=1,timeout=0,minproto=5,maxproto=5,direct,pipe_ino=10489)
gfs1.iad-aws.prod.sli.io:/sec-gfs1 on /mnt/sec-gfs1 type fuse.glusterfs (ro,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
gfs1.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (ro,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If any host will not return all expected mounts as an output, mount volumes manually:
[local]% ssh <IP address of the host> "mount /mnt/gfs1; mount /mnt/sec-gfs1"
If mounting unsuccessful, escalate to BU SaaS.
[X] 2.7.1.2. Ensure mounts are healthy
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo ${host}; ssh ${host} "ls /mnt/sec-gfs1/_gfs_healthcheck_ 1>/dev/null"; done
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo ${host}; ssh ${host} "ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null"; done
All hosts should produce no output.
If nothing printed out, proceed to the next step.
If any host will produce output like following:
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
or
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
or
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: Transport endpoint is not connected
or
ls: cannot access /mnt/sec-gfs1/_gfs_healthcheck_: No such file or directory
Remount volumes on a specific hosts:
[local]% ssh <IP address of the host> "umount /mnt/gfs1; umount /mnt/sec-gfs1; sync; sleep 2; mount /mnt/gfs1; mount /mnt/sec-gfs1"
Recheck mounts starting from 2.7.1.1.
[if there are issues after second run, escalate to BU SaaS]
[X] 2.7.1.3. Ensure Apache is running
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo ${host}; ssh ${host} "systemctl status httpd -l"; done
on any problem, restart apache on a respective host:
[local]% ssh <IP address of the host> "systemctl restart httpd"
[X] 2.7.1.3. Ensure Localbrain is running
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo ${host}; ssh ${host} "systemctl status braind -l"; done
on any problem, restart apache on a respective host:
[local]% ssh <IP address of the host> "systemctl stop braind; sleep 30; systemctl start braind"
2.7.2. Dory servers
[X] 2.7.2.1. Ensure mounts
[local]% for host in 10.79.114.251 10.79.114.88; do echo ${host}; ssh ${host} "mount | grep gfs1"; done
all hosts should report gfs1 is mounted
gfs1.iad-aws.prod.sli.io:/gfs1 on /mnt/gfs1 type fuse.glusterfs (ro,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
If any host will not return all expected mounts as an output, mount volumes manually:
[local]% ssh <IP address of the host> "mount /mnt/gfs1"
If mounting unsuccessful, escalate to BU SaaS.
[X] 2.7.2.2. Ensure mounts are healthy
[local]% for host in 10.79.114.251 10.79.114.88; do echo ${host}; ssh ${host} "ls /mnt/gfs1/_gfs_healthcheck_ 1>/dev/null"; done
All hosts should produce no output.
If nothing printed out, proceed to the next step.
If any host will produce output like following:
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: Transport endpoint is not connected
or
ls: cannot access /mnt/gfs1/_gfs_healthcheck_: No such file or directory
Remount volumes on a specific hosts:
[local]% ssh <IP address of the host> "umount /mnt/gfs1; sync; sleep 2; mount /mnt/gfs1"
Recheck mounts starting from 2.7.2.1.
[if there are issues after second run, escalate to BU SaaS]
[X] 2.7.2.3. Ensure Apache is running
[local]% for host in 10.79.114.251 10.79.114.88; do echo ${host}; ssh ${host} "systemctl status httpd -l"; done
on any problem, restart apache on a respective host:
[local]% ssh <IP address of the host> "systemctl restart httpd"
[X] 2.7.2.3. Ensure Dory is running
[local]% for host in 10.79.114.251 10.79.114.88; do echo ${host}; ssh ${host} "systemctl status dory -l"; done
on any problem, restart apache on a respective host:
[local]% ssh <IP address of the host> "systemctl stop dory; sleep 30; systemctl start dory"
[X] 2.7.3. Ensure servers return valid information instead of errors.
Use Nagios (https://nagios-master.sli.io/nagios/) monitoring to check for errors. If nagios reports issues like following (errors 400 or 404) for any IAD cluster servers, escalate to BU SaaS
search1-3.iad.prod asahi braindead
Notifications for this service have been disabled
CRITICAL 02-08-2021 22:45:42 5d 21h 24m 45s 3/3 asahi/search1-3.iad-aws.prod.sli.io returned 400
```
### Following steps hasn't been executed because cluster is not in production now and we don't want ELB to send any production load to this cluster.
```
If any of the following steps fail, escalate to BU SaaS.
[ ] 2.8. Returning Cluster 1 (IAD) back to rotation:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do ssh ${host} 'runuser -l gbrain bash -c "webcontrol vip insert"'; done
(allow a minute before performing next step)
2.9. Ensure Cluster 1 (IAD) is serving to the clients
[ ] 2.9.1. Check vip status on searches:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do ssh ${host} 'runuser -l gbrain bash -c "webcontrol vip status"'; done
[ ] 2.9.2. Request healthcheck document from the search hosts:
[local]% for host in 10.79.114.86 10.79.114.150 10.79.114.92 10.79.114.15 10.79.114.94 10.79.114.65; do echo -n "${host}: "; curl -A "Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:59.0) Gecko/20100101 Firefox/59.0" http://${host}/shared/vip-status.txt; done
expecting all hosts to return UP
```
Sign in with Wallet
Connect another wallet