#Failed to create a file with surrogate pair characters

1 messages · Page 1 of 1 (latest)

lethal sapphire
#

I receive the following event constantly and wonder if anyone knows how to see what file contains the bad characters?

ERROR wafl.dir.surrpair.filename: Failed to create a file with surrogate pair characters in the name in the directory /vol/<path>

summer totem
#

I’ve often seen that when something tries to make a file with “/“ in the name vs path

#

Or “\”? One of the two

rigid locust
#

you can simply change the volume language tu utf8mb4 which lets you create files with non-BMP codepoints in their filename.

wide pollen
#

Here's some KBs that might help:
https://kb.netapp.com/onprem/ontap/hardware/wafl_dir_surrpair_filename_in_EMS_for_volume_with_utf8_language_Active_IQ_Wellness_Incident

https://kb.netapp.com/onprem/ontap/da/NAS/Unable_to_create_files_Error_STATUS_OBJECT_NAME_INVALID_0xc0000033_Why_cant_I_use_supplementary_characters_in_my_filename#

I believe if you look in your EMS logs, it will list the directory/files that are having the issue.

@rigid locust I tried to change the language on a volume on a cluster that is running 9.8P18. When I use "volume language -vserver vs01 -volume vol1 -language utf8mb4" I get a response "There are no entries matching your query". If I try "volume modify -vserver vs01 -volume vol1 -language utf8mb4" it resonse with "Error: command failed: Changing the language of a volume is not supported. Contact technical support for assistance"

Is this just not supported in 9.8? If not, what code level is it supported in?

rigid locust
wide pollen
#

Cool, that did it. Thanks, @rigid locust.
An important thing to note is that once you've converted to utf8mb4, it will not let you go back.

Are there any downsides to changing this? It gives a warning not to do it unless directed to by NetApp support. That makes me worry about doing it on a production volume. Especially when all our volumes are standardized on C.UTF-8.

rigid locust
#

as the KB article states, there are no reported issues (and I can't think of any, given that what it stores is always UTF-8, it's just that this now allows 4-byte sequences instead of only 3-byte sequences)

lethal sapphire
#

Thanks for the replies but if I read between the lines here it seems like there's no way to see what file it complains about?
Changing language of the volume is not something we wish to consider at this time.

rigid locust
#

I don't think you can see which file the client is trying to create, no. You would certainly see an error on the client though, but if users don't complain about them not being able to save a file as "😃 .txt" you're out of luck. I'm not even sure you would see anything in the CIFS audit log either

lethal sapphire