problems deleting with 2.3S1 and more than one disk

From: Raul Rivero <rivero@dont-contact.us>
Date: Sat, 05 Feb 2000 16:49:34 +0100

Hi all,

Two weeks ago we upgraded our proxy/cache farm from
squid 2.2Stable5 to 2.3Stable1 and we have problems
with one of our caches which have three disks.

When we installed that machine we had two options:
define an unique RAID0 "disk"; or keep splited the
three disks and define three "cache_dir".

Well, we choose second option so now we have ...

replacement_policy GDSF
cache_swap_low 90
cache_swap_high 95
cache_dir ufs /usr/local/SI/squid2/cache/disco1 7000 16 256
cache_dir ufs /usr/local/SI/squid2/cache/disco2 7000 16 256
cache_dir ufs /usr/local/SI/squid2/cache/disco3 8000 16 256

... but a "df -k" says ...

/dev/sdb1 8746648 8064788 237548 97%
/usr/local/SI/squid-2.2/cache/disco1
/dev/sdc1 8746648 8074712 227624 97%
/usr/local/SI/squid-2.2/cache/disco2
/dev/hdc1 9845200 9224208 120876 99%
/usr/local/SI/squid-2.2/cache/disco3
                                      ^^

We have machines with only one "cache_dir", and everything
is working fine with these ones. So, we think that squid
have problems purging objects with more that one "cache_dir".

Any idea?

--
Raul Rivero                           | Univ. de Oviedo               |
rivero@si.uniovi.es                   | Edif. Historico - Aula Clarin |
http://www.uniovi.es/~rivero/         | San Francisco, 1              |
Tel: +34 985102908 Fax: +34 985102915 | 33003 - Oviedo (Spain)        |
Received on Sat Feb 05 2000 - 15:13:23 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:58 MST