[ARISTA] adding 7060_cs32s to eMMC exclusions#2982
Conversation
Following PR 2774 we added the 7060-cx32s according to the guidelines of PR 2780 This adds the 7060-cx32s to the list f devices that mount /var/log as a tmpfs to mitigate eMMC wearout Signed-off-by: Michel Moriniaux <m.moriniaux@criteo.com>
|
@MichelMoriniaux I remember we had a chat about this change on my PR. You mentioned that you could keep it internal. Did things change since then? We don't really see 7060 having hard drive issue and we have concerns for losing logs across reboot. If this is not necessary, we would like to keep this change out. Can you let us know:
Thanks, |
|
@Staphylo can you please comment, as you suggested to us to implement this As usual we prefer not to keep code internal that can be of utility to the community, especially code that can prevent bricking devices. We had agreed to remove our PR in favor of yours and keep it internal until we amended your code to include this platform. As we are moving our patches to 201904, this came up. As you remember our original implementation was not vendor specific but depended on detecting the presence of an eMMC kernel driver, we believe that eMMC will fail at one point ( even if it's after 10 years ) and that it is best practice to have logs be streamed out of the device as soon as possible. |
|
We indeed recommended to add the 7060CX-32S in the list of products that log in a tmpfs rather than the actual flash device. Log writing can be IO intensive depending on the verbosity. |
Following PR 2774 we added the 7060-cx32s according to the guidelines of PR 2780 This adds the 7060-cx32s to the list f devices that mount /var/log as a tmpfs to mitigate eMMC wearout Signed-off-by: Michel Moriniaux <m.moriniaux@criteo.com>
|
retest vs please |
Following PR 2774 we added the 7060-cx32s according to the guidelines of
PR 2780
This adds the 7060-cx32s to the list f devices that mount /var/log as a
tmpfs to mitigate eMMC wearout
Signed-off-by: Michel Moriniaux m.moriniaux@criteo.com