Running SAIO config generating recon cache errors

Asked by Chris Almond on 2012-07-13

I've got a running SAIO system extended to use swauth, https & Swift3. I noticed the log files /var/log/swift/storageX.log and storage.error were growing quit large. Every few seconds the following would appear in the error log:

Jul 13 09:48:36 swift-saio-0 object-replicator Exception dumping recon cache: #012Traceback (most recent call last):#012 File "/home/saio/swift/swift/common/utils.py", line 1196, in dump_recon_cache#012 with lock_file(cache_file, lock_timeout, unlink=False) as cf:#012 File "/usr/lib/python2.6/contextlib.py", line 16, in __enter__#012 return self.gen.next()#012 File "/home/saio/swift/swift/common/utils.py", line 761, in lock_file#012 fd = os.open(filename, flags)#012OSError: [Errno 13] Permission denied: '/var/cache/swift/object.recon'
Jul 13 09:49:41 swift-saio-0 object-replicator last message repeated 2 times

The /var/cache/swift directory did not exist. After creating it and chowning it to my swift process UID/GID ("saio") the log files stopped growing. And, the following files were immediately created inside /var/cache/swift:

-rw------- 1 saio saio 270 2012-07-13 09:49 account.recon
-rw------- 1 saio saio 272 2012-07-13 09:49 container.recon
-rw------- 1 saio saio 52 2012-07-13 09:49 obje ct.recon

Note - I don't have any pipeline or filter settings for recon in any of my account/object/container server conf files. Looks like the SAIO config documentation should at a minimum call out creation of the /var/cache/swift directory to prevent log files from blowing up. (or add recon config to all of the conf files?)

Question information

Language:
English Edit question
Status:
Answered
For:
OpenStack Object Storage (swift) Edit question
Assignee:
No assignee Edit question
Last query:
2012-07-13
Last reply:
2012-09-23
Tom Fifield (fifieldt) said : #1

Thanks for the report - as per the bug https://bugs.launchpad.net/swift/+bug/1024477 this is now fixed, so marking the question as answered.b

Can you help with this problem?

Provide an answer of your own, or ask Chris Almond for more information if necessary.

To post a message you must log in.