Greetings list,
I have a small XRootD v5.3.1 cluster (3 data servers and 1 redirector VM) where some files have gone missing recently. I think there was an OOM on a storage server last week (not XRootD's fault) that caused various things to crash. Either way, we're in this
scenario now where the files don't seem to exist on disk but XRootD still has metadata for them. I don't know how to delete these files and could use some help!
When I ask the server for the file, I see:
# xrdfs root://uct3-xrd.mwt2.org:1094 ls /atlas/uct3/data/users/jano/qual/matched/truth
[!] Some of the requests failed. The result may be incomplete.
/atlas/uct3/data/users/jano/qual/matched/truth/ntuple_0.root
Locating the file, I see:
# xrdfs root://uct3-xrd.mwt2.org:1094 locate /atlas/uct3/data/users/jano/qual/matched/truth/ntuple_0.root
[::192.170.227.124]:1094 ServerPending ReadWrite
If I use the '-r' flag to ignore caches, I get:
# xrdfs root://uct3-xrd.mwt2.org:1094 locate -r /atlas/uct3/data/users/jano/qual/matched/truth/ntuple_0.root
[ERROR] Server responded with an error: [3011] No servers have the file
Now I'm not sure how to proceed. I can't rm the file from the xrdfs interface:
[root@uct3-s1 ~]# xrdfs root://uct3-xrd.mwt2.org:1094 rm /atlas/uct3/data/users/jano/qual/matched/truth/ntuple_0.root
[ERROR] Server responded with an error: [3011] Unable to access file; eligible servers shunned.
These files can be regenerated - so dropping them is no problem. However I'm not sure how to get XRootD to drop this ghost (and others).
Any insight would be appreciated.
Thanks,
Lincoln
Use REPLY-ALL to reply to list
To unsubscribe from the XROOTD-L list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1