Error Message

Nov 22 08:21:35 server101 scini[6049]: Error (sftp): File /etc/emc/scaleio/scini_sync/scini_key must exist.
Nov 22 08:21:35 server101 scini[6051]: driver_sync.sh END - Wed Nov 22 08:21:35 EST 2023
Nov 22 08:21:35 server101 scini[5193]: driver_sync failed! Please check driver_sync.log
Nov 22 08:21:35 server101 scini[5193]: No match after driver_sync fetch attempt.
Nov 22 08:21:35 server101 scini[5193]: Attempting to find a best match...
Nov 22 08:21:35 server101 systemd[1]: Started Automounts filesystems on demand.

..

Nov 22 08:21:38 server101 vxvm-boot[18297]: Starting up VxVM:
...
Nov 22 08:21:39 server101 kernel: PowerFlex drvMain_Init:61 :Driver starting load process. HZ 1000...
Nov 22 08:21:39 server101 kernel: PowerFlex mosCpu_Init:80 :eax=00000960, ebx=00000dac, ecx=00000064, edx=00000000
Nov 22 08:21:39 server101 kernel: PowerFlex mosT10Dif_Init:933 :(T10DIF) T10DIF layer using CPU (PCLMUL) acceleration !!!
Nov 22 08:21:39 server101 kernel: PowerFlex drvMain_Init:63 :Load process done. Status SUCCESS
Nov 22 08:21:39 server101 scini[5163]: Success loading module /bin/emc/scaleio/scini.ko
Nov 22 08:21:39 server101 scini[5163]: Found device special file /dev/scini
Nov 22 08:21:39 server101 kernel: VxVM vxio V-5-0-1990 driver version 7.4.2.4300 installed
Nov 22 08:21:39 server101 kernel: VxVM vxspec V-5-0-1990 driver version 7.4.2.4300 installed
Nov 22 08:21:39 server101 vxvm-boot[30112]: cat: /proc/sys/kernel/hotplug: No such file or directory
Nov 22 08:21:39 server101 vxvm-boot[21946]: VxVM general startup...
Nov 22 08:21:39 server101 kernel: PowerFlex mapMdmLayer_SetIniGUID:6107 :No such SDC instance, allocate one..
Nov 22 08:21:39 server101 kernel: PowerFlex ioctl_HandleSetIniGuid:644 :Set GUID (35908BB8-9DD7-436D-B326-150093745E95) finished . status 65 (SUCCESS)
Nov 22 08:21:39 server101 scini[28737]: Loading MDMs from configuration file /etc/emc/scaleio/drv_cfg.txt
Nov 22 08:21:39 server101 scini[28737]: Loading MDMs from configuration file /etc/emc/scaleio/drv_cfg.txt
Nov 22 08:21:39 server101 scini[28737]: Successfully set GUID 35908BB8-9DD7-436D-B326-150093745E95
Nov 22 08:21:39 server101 scini[28737]: Successfully add mdm from cfg line 10
Nov 22 08:21:39 server101 kernel: PowerFlex ioctl_HandleAddMdm:469 :Create finish for MDM. First I

...

Nov 22 08:21:44 server101 scini[5163]: scinia is not ready yet...

The ScaleIO daemon (scini) has not finished starting and creating the storage before the VxVM process completes and comes online. VxFEN will then try to start and fail due to the missing coordinator disks.

This issue has been fixed in the following patches:

•    For InfoScale 7.4.2 RHEL8 a private hotfix is available ( vm-rhel8_x86_64-HotFix-7.4.2.5302 ) and the fix will be included in the forthcoming update 8 patch. 
•    For InfoScale 8.0.2 the fix is included in the update4 patch https://downloads.infoscale.com/infoscale/REL333811/8.0.2.2400?q=UPD95&fileNumber=FILE634756&updateNumber=UPD954051

•    For InfoScale 8.0 the fix is included in the update3 patch https://downloads.infoscale.com/infoscale/REL150377/8.0.0.3300?q=UPD82&fileNumber=FILE345774&updateNumber=UPD826886

It is recommended that the above patches be installed. If the patches are not installed, then the following workaround will need to be implemented.
 

1. Stop vxfen.service.

# systemctl stop vxfen.service

2. Edit the file /usr/lib/systemd/system/vxfen.service and add the scini.service to the After attribute. i.e. 

After=rsyslog.service veki.service llt.service gab.service sdcd.service scini.service

3. Copy the sample script of pre_vxfenstart.

# cp /opt/VRTSvcs/vxfen/sample_scripts/pre_vxfenstart /opt/VRTSvcs/vxfen/bin/
# chmod a+x /opt/VRTSvcs/vxfen/bin/pre_vxfenstart

4. Modify the script /opt/VRTSvcs/vxfen/bin/pre_vxfenstart appending the following lines to the end of the file

LC_ALL=C $VXDISK scandisks 2>/dev/null
LC_ALL=C $VXDISK -oalldgs list >$VCS_LOG_TMP/$VXFEN_TEMP_PRE.scandisks.$$ 2>/dev/null

RETVAL=0
return $RETVAL

5. Restart vxfen.service or reboot.

# systemctl start vxfen.service

After rebooting a node, the VxFEN service fails to start when using EMC ScaleIO. This can cause the node to fail to join the cluster.

ETrack: 4153457 JIRA: STESC-8504