-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 4/10/2014 1:47 a.m., Stephen Baynes wrote: > At times we automatically start Squid from a clean setup. We run > squid -z to create the cache directories and then start Squid for > normal operation. The question is how to know when squid -z has > finished and it is safe to start Squid for normal operation. > > This is not a problem when one has one squid worker, one uses the > -N (don't background) option as well and just waits for the process > to finish. > > However we want to use Squid with multiple workers and the problem > is the -N option overrides the number of workers and it sets up > only one cache directory. (Squid has one cache directory per > worker.) If one does not wait then when Squid starts for normal > operation the cache directories are not ready and it fails. > > One can't even use the PID file to find out a process to wait for, > as it seems that it does not even write a PID file when the -z > option is used. > > Any ideas on how to find out when it is finished creating the cache > directories? This appears to be one of the little details that have not been polished yet with SMP. What I have been doing for this type of check is looking in /dev/shm (on Linux, may not be there on other OS) to see if the SMP sockets are still active. Squid removes the SHM device "files" when its done. Amos -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (MingW32) iQEcBAEBAgAGBQJULvRvAAoJELJo5wb/XPRj7/8H/iNVD3ND5TmZIhk2MtzX3Vg/ fhfeYEQnAGBAcKgtVwVNPihI6EsVktZt9OcNWajQkkv8hPLSQ4IQBVSI4NgXzAa9 6tYhEtSydoYjWevjKcwf0Wyea6TXIcULUFEz5dVsu8ISflHisDdqYAgQZM691RCD p8P+W0wdPeF7KIqX214aADHr1m/t0hfjkEHW5M5KRIN6O1IN4wOg38H5u3qhJM/z j5qgcoNL+8pKGA4dasurQRe8yVCgeNVQp9ipFZw7U0oQ3+oVXH2n+9MWYP+QaZMi nLipoGEkj5BNd1bEvklArWUyQBcbUb64bCbTXQl8Dhse8qthWQX7LhHWXNZbOoI= =eEx1 -----END PGP SIGNATURE----- _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
