Forum Discussion

ignes_ingram's avatar
ignes_ingram
Novice I
3 years ago

Hi. Anybody seen this error on vVols? 20T1009.319Z

Hi. Anybody seen this error on vVols? 20T1009.319Z cpu24:2097621)ScsiDevice: 3791: Handle REPORTED LUNS CHANGED DATA unit attention 2023-01-20T1009.319Z cpu47:2111187)VMW_SATP_ALUA: satp_alua_issueCommandOnPath Path (vmhba4T1:L249) command 0xa3 failed with transient error status Transient storage condition, suggest retry. sense data: 0x6 0x3f 0xe. Waiting for 20 seconds

6 Replies

  • So that isn't a real failure or issue. Whenever volumes are connected or disconnected to a host on the array, the array will issue a Unit Attention (UA). What does the unit attention do? Short answer it let's the host/init know that there has been a change to the scsi devices connected to the host and that it should refresh/update the device listings. This means that you don't need to rescan from the host every time a volume is connected or disconnected from the host in order to get the scsi device list to update. This is very important for vvols as many volume connect and disconnect operations are going to happen and we don't want the host to have to manually rescan storage each time.
  • All that log line is saying is that the scsi op 0xa3 failed because a unit attention was issued. The host will then retry that scsi op after it updates it's device listing.
  • Thanks for the feedback alex.carver. Could a too aggressive DRS priority cause these messages to be logged too frequently?
  • Ya, but it's not really a "too frequently" issue. It's a matter more of something getting logged in general. vMotions, power on, power off, clone, or even just browsing the vvol datastore will cause these messages to pop up. Essentially any operation that causes a volume to be connected or disconnected from a host will trigger a UA
  • Hi. To add to this. The customer has RecoverPoint in play. The attached errors pop up at random. This is not isolated to a specific VM. It will clear then appear again. Any ideas?
  • I would really hope that recover point wouldn't be throwing those errors because of a unit attention being triggered for volumes on the array. That could be a bug on their end, but I really hope it's not the case that when a 0x89 (ats heartbeat) fails and has to be retry due to a UA that recover point then fails.