HP StoreAll OS 6.5.1 Release Notes
Query database will be incomplete during this time. HP recommends that you not use the database
until the Offline Synchronizer has completed.
• If an Express Query-enabled file system has a continuously high work load, it is possible that the
freshness of the Express Query data could be affected.
To handle this issue, consider the following:
◦ Stop the client work load until all files can be updated in the Express Query database.
◦ Periodically check the events log for file system notifications about space being consumed.
◦ Query the freshness of the Express Query data as shown in the following example:
# curl –gi
http://<ip>/<share>/<file_name>?attributes=system::size&query=system::size>2048&freshness
HTTP/1.1 200 OK
Date: Tue, 07 Jan 2014 15:39:46 GMT
Server: Apache/2.2.14 (Unix) mod_ssl/2.2.14 OpenSSL/0.9.8n DAV/2
Api-Version: 2
Freshness-Complete-Date: Tue, 07 Jan 2014 08:16:18 MST
Freshness-Partial-Date: Tue, 07 Jan 2014 08:16:19 MST
Content-Type: application/json
Content-Length: 63
In this example, the Freshness-Complete-Date value indicates the timestamp of file system
operations already processed by Express Query. Other file system operations that occurred
after this timestamp might not have been processed by Express Query yet.
• Large file systems enabled for Express Query and configured for custom metadata can cause long
response times for queries. For example, you have a file system with 100,000 files and four
custom attributes per file. If you run a query on this file system (using the REST API GET request
command), it may take up to 20 minutes to receive a response.
To avoid this issue:
◦ Avoid queries that have a large result set.
◦ Avoid queries that contain system and custom metadata. Queries with custom metadata
perform better when system metadata (except for path name and POID) are not selected in
the result set.
• If you have an Express Query-enabled file system that contains tens of millions of files and files
are renamed frequently (approximately 1 million files per hour), this may cause Express Query
to fail with the following error posted in the ajdatabase.log file (located in /usr/local/
ibrix/logs):
./daily/ajdatabase.log.3.gz:ERROR: [mb_gen] [22619] [47641778470832]
[2014-01-08 01:59:19.536] About to close 12 because of packet exceeded:
1158659156
Although Express Query will attempt to restart, the issue will occur again and be repeated in a
continuous loop.
To resolve the issue, complete the following steps:
NOTE: In this procedure “ifs1” is the file system name. Be sure to replace any reference to
“ifs1” with the applicable file system name.
1. Verify the current state of the file system by executing the following command:
ibrix_archiving -1
Command output example:
FS NAME STATE
------- -----
ifs1 AbnormalDatabase(DBNotAdded)
2. Force the Manual Intervention Failure (MIF) state by executing the following command:
16 Workarounds