Aleph v20 "No space left on device" ; 100s of DNS queries/second; ue_23

• Article Type: General
• Product: Aleph
• Product Version: 20

Description:
After upgrading to Aleph v20 we are getting this error logging on to the server:

cp: writing `/exlibris/aleph/a20_2/tmp/f_symbol.22502': No space left on device

At another site, the problem manifested itself as follows:

Our network group says that we are experiencing 80 to 100 queries per second on our DNS server from Aleph v. 20 which is running on our test server (this stops when Aleph is shutdown and starts right away when Aleph is started back up).

For example:
14-May-2010 09:35:15.301 client 136.167.2.151#57748: query: . IN A +
14-May-2010 09:35:15.301 client 136.167.2.151#57749: query: 0.0.0.0.in-addr.arpa IN PTR +

Resolution:
Using the methods described in Article 000033291 ("/exlibris filesystem 100% full; locating files to delete **MASTER RECORD**"), I found that there was a 102 gig file in the abc50 $data_scratch:

-rw-rw-r-- 1 aleph exlibris 102579322880 May 25 11:11 run_e_23.24008

This file is the log of the ue_23 process which had the following messages written more than 100 times every second:

2010-05-04 07:28:21 [inf] HK system not available.
can't get host by address or by name2010-05-04 07:28:21 Error: Unable to get host by name

util c/1 showed that the current ue_23 had 17,398 cpu seconds.

This process does not need to run at your site and should not be running. I have stopped it, but, in order to prevent it from starting automatically the next time Aleph is restarted, see Article 000043557 ("What is ue_23? Preventing it from starting automatically").