[lvm-devel] master - lvmetad: Wait at least 80s for the initial scan.
Hannes Reinecke
hare at suse.de
Tue Mar 25 16:00:46 UTC 2014
On 03/25/2014 08:39 AM, Petr Rockai wrote:
> Gitweb: http://git.fedorahosted.org/git/?p=lvm2.git;a=commitdiff;h=c13f5dbb2521901f723d4c40fd5992ee0b992d8d
> Commit: c13f5dbb2521901f723d4c40fd5992ee0b992d8d
> Parent: 65bbfdf74ddbb16579c4fa5fc76e1f51aee0da22
> Author: Petr Rockai <prockai at redhat.com>
> AuthorDate: Tue Mar 25 15:53:36 2014 +0100
> Committer: Petr Rockai <prockai at redhat.com>
> CommitterDate: Tue Mar 25 16:38:35 2014 +0100
>
> lvmetad: Wait at least 80s for the initial scan.
>
> ---
> lib/cache/lvmetad.c | 14 +++++++++-----
> 1 files changed, 9 insertions(+), 5 deletions(-)
>
> diff --git a/lib/cache/lvmetad.c b/lib/cache/lvmetad.c
> index 040a790..3ee850c 100644
> --- a/lib/cache/lvmetad.c
> +++ b/lib/cache/lvmetad.c
> @@ -142,6 +142,7 @@ static daemon_reply _lvmetad_send(const char *id, ...)
> daemon_reply repl;
> daemon_request req;
> int try = 0;
> + int time = 0, wait;
>
> retry:
> req = daemon_request_make(id);
> @@ -158,7 +159,7 @@ retry:
> daemon_request_destroy(req);
>
> if (!repl.error && !strcmp(daemon_reply_str(repl, "response", ""), "token_mismatch") &&
> - try < 60 && !test_mode()) {
> + try < 10 && time < 80000000 && !test_mode()) {
> /*
> * If another process is trying to scan, they might have the
> * same future token id and it's better to wait and avoid doing
> @@ -171,12 +172,15 @@ retry:
> * the update, we back off for a short while (0.2-2 seconds) and
> * try again.
> */
> - if (!strcmp(daemon_reply_str(repl, "expected", ""), "update in progress") || try % 5)
> - usleep( 50000 + random() % 450000 ); /* 0.05 - 0.5s */
> - else
> + if (!strcmp(daemon_reply_str(repl, "expected", ""), "update in progress") || try % 5) {
> + wait = 50000 + random() % 450000; /* 0.05 - 0.5s */
> + time += wait;
> + usleep( wait );
> + } else {
> /* If the re-scan fails here, we try again later. */
> lvmetad_pvscan_all_devs(_lvmetad_cmd, NULL);
> - ++ try;
> + ++ try;
> + }
> daemon_reply_destroy(repl);
> goto retry;
> }
>
Can't we have a configuration setting for this?
I can imagine plenty of scenarios where the default value might not be
applicable, being either to short or too long.
Cheers,
Hannes
--
Dr. Hannes Reinecke zSeries & Storage
hare at suse.de +49 911 74053 688
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg
GF: J. Hawn, J. Guild, F. Imendörffer, HRB 16746 (AG Nürnberg)
More information about the lvm-devel
mailing list