From: rpeterso@sourceware.org <rpeterso@sourceware.org>
To: cluster-devel.redhat.com
Subject: [Cluster-devel] cluster/gfs/gfs_fsck fs_recovery.c fsck.h init ...
Date: 14 Jun 2006 19:12:09 -0000 [thread overview]
Message-ID: <20060614191209.5876.qmail@sourceware.org> (raw)
CVSROOT: /cvs/cluster
Module name: cluster
Changes by: rpeterso at sourceware.org 2006-06-14 19:12:07
Modified files:
gfs/gfs_fsck : fs_recovery.c fsck.h initialize.c log.c main.c
pass1.c pass2.c pass3.c pass4.c pass5.c super.c
util.c util.h
Log message:
This addresses bugzilla bug #156009 - gfs fsck needs a good
review of logging. I improved logging and added some "warm
fuzzy" feel-good messages to let the users know it's not hung.
Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/fs_recovery.c.diff?cvsroot=cluster&r1=1.3&r2=1.4
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/fsck.h.diff?cvsroot=cluster&r1=1.3&r2=1.4
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/initialize.c.diff?cvsroot=cluster&r1=1.10&r2=1.11
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/log.c.diff?cvsroot=cluster&r1=1.4&r2=1.5
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/main.c.diff?cvsroot=cluster&r1=1.3&r2=1.4
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/pass1.c.diff?cvsroot=cluster&r1=1.13&r2=1.14
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/pass2.c.diff?cvsroot=cluster&r1=1.11&r2=1.12
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/pass3.c.diff?cvsroot=cluster&r1=1.8&r2=1.9
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/pass4.c.diff?cvsroot=cluster&r1=1.11&r2=1.12
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/pass5.c.diff?cvsroot=cluster&r1=1.9&r2=1.10
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/super.c.diff?cvsroot=cluster&r1=1.6&r2=1.7
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/util.c.diff?cvsroot=cluster&r1=1.6&r2=1.7
http://sourceware.org/cgi-bin/cvsweb.cgi/cluster/gfs/gfs_fsck/util.h.diff?cvsroot=cluster&r1=1.3&r2=1.4
--- cluster/gfs/gfs_fsck/fs_recovery.c 2005/08/01 16:20:48 1.3
+++ cluster/gfs/gfs_fsck/fs_recovery.c 2006/06/14 19:12:07 1.4
@@ -75,15 +75,15 @@
* Returns: 0 on success, -1 on failure
*/
int reconstruct_journals(struct fsck_sb *sdp){
- int i;
+ int i;
- log_warn("Clearing journals (this may take a while)\n");
- for(i=0; i < sdp->journals; i++) {
- if((i % 10) == 0)
- log_at_notice(".");
- if(reconstruct_single_journal(sdp, i))
- return -1;
- }
- log_notice("Cleared journals\n");
- return 0;
+ log_notice("Clearing journals (this may take a while)");
+ for(i=0; i < sdp->journals; i++) {
+ if((i % 2) == 0)
+ log_at_notice(".");
+ if(reconstruct_single_journal(sdp, i))
+ return -1;
+ }
+ log_notice("\nJournals cleared.\n");
+ return 0;
}
--- cluster/gfs/gfs_fsck/fsck.h 2005/08/01 16:20:48 1.3
+++ cluster/gfs/gfs_fsck/fsck.h 2006/06/14 19:12:07 1.4
@@ -26,7 +26,7 @@
int no:1;
};
-
+extern uint64_t last_fs_block;
int initialize(struct fsck_sb *sbp);
void destroy(struct fsck_sb *sbp);
--- cluster/gfs/gfs_fsck/initialize.c 2005/12/15 23:22:22 1.10
+++ cluster/gfs/gfs_fsck/initialize.c 2006/06/14 19:12:07 1.11
@@ -98,6 +98,7 @@
{
uint32_t i;
+ log_info("Freeing buffers.\n");
if(sdp->riinode){
free(sdp->riinode);
sdp->riinode = NULL;
@@ -108,7 +109,7 @@
}
if(sdp->rooti){
free(sdp->rooti);
- sdp->rooti=NULL;
+ sdp->rooti = NULL;
}
if(sdp->jindex){
@@ -196,6 +197,7 @@
rmin = ri->ri_data1;
}
+ last_fs_block = rmax;
for (i = 0; i < sdp->journals; i++)
{
@@ -406,6 +408,7 @@
log_warn("Unable to unblock other mounters - manual intevention required\n");
log_warn("Use 'gfs_tool sb <device> proto' to fix\n");
}
+ log_info("Syncing the device.\n");
fsync(sbp->diskfd);
}
empty_super_block(sbp);
--- cluster/gfs/gfs_fsck/log.c 2005/08/01 16:20:48 1.4
+++ cluster/gfs/gfs_fsck/log.c 2006/06/14 19:12:07 1.5
@@ -45,11 +45,13 @@
case MSG_DEBUG:
printf("(%s:%d)\t", file, line);
vprintf(format, args);
+ fflush(stdout);
break;
case MSG_INFO:
case MSG_NOTICE:
case MSG_WARN:
vprintf(format, args);
+ fflush(stdout);
break;
case MSG_ERROR:
case MSG_CRITICAL:
--- cluster/gfs/gfs_fsck/main.c 2005/02/22 21:57:54 1.3
+++ cluster/gfs/gfs_fsck/main.c 2006/06/14 19:12:07 1.4
@@ -20,6 +20,8 @@
#include "fsck.h"
#include "log.h"
+uint64_t last_fs_block;
+
void print_map(struct block_list *il, int count)
{
int i, j;
@@ -128,40 +130,41 @@
log_notice("Starting pass1\n");
if (pass1(sbp))
return 1;
- log_notice("Pass1 complete\n");
+ log_notice("Pass1 complete \n");
log_notice("Starting pass1b\n");
if(pass1b(sbp))
return 1;
- log_notice("Pass1b complete\n");
+ log_notice("Pass1b complete \n");
log_notice("Starting pass1c\n");
if(pass1c(sbp))
return 1;
- log_notice("Pass1c complete\n");
+ log_notice("Pass1c complete \n");
log_notice("Starting pass2\n");
if (pass2(sbp, &opts))
return 1;
- log_notice("Pass2 complete\n");
+ log_notice("Pass2 complete \n");
log_notice("Starting pass3\n");
if (pass3(sbp, &opts))
return 1;
- log_notice("Pass3 complete\n");
+ log_notice("Pass3 complete \n");
log_notice("Starting pass4\n");
if (pass4(sbp, &opts))
return 1;
- log_notice("Pass4 complete\n");
+ log_notice("Pass4 complete \n");
log_notice("Starting pass5\n");
if (pass5(sbp, &opts))
return 1;
- log_notice("Pass5 complete\n");
+ log_notice("Pass5 complete \n");
/* print_map(sbp->bl, sbp->last_fs_block); */
+ log_notice("Writing changes to disk\n");
destroy(sbp);
return 0;
--- cluster/gfs/gfs_fsck/pass1.c 2006/04/20 20:37:50 1.13
+++ cluster/gfs/gfs_fsck/pass1.c 2006/06/14 19:12:07 1.14
@@ -883,7 +883,8 @@
stack;
return -1;
}
- log_debug("RG at %"PRIu64" is %u long\n", rgd->rd_ri.ri_addr, rgd->rd_ri.ri_length);
+ log_debug("RG at %"PRIu64" is %u long\n", rgd->rd_ri.ri_addr,
+ rgd->rd_ri.ri_length);
for (i = 0; i < rgd->rd_ri.ri_length; i++) {
if(block_set(sbp->bl, rgd->rd_ri.ri_addr + i,
meta_other)){
@@ -903,7 +904,7 @@
if(next_rg_meta_free(rgd, &block, first, &mfree))
break;
-
+ warm_fuzzy_stuff(block);
if(get_and_read_buf(sbp, block, &bh, 0)){
stack;
log_crit("Unable to retrieve block %"PRIu64
--- cluster/gfs/gfs_fsck/pass2.c 2005/03/07 22:04:58 1.11
+++ cluster/gfs/gfs_fsck/pass2.c 2006/06/14 19:12:07 1.12
@@ -808,6 +808,7 @@
}
int error = 0;
+ log_info("Checking directory inodes.\n");
/* Grab each directory inode, and run checks on it */
for(i = 0; i < sbp->last_fs_block; i++) {
@@ -824,7 +825,7 @@
if(q.block_type != inode_dir)
continue;
- log_info("Checking directory inode at block %"PRIu64"\n", i);
+ log_debug("Checking directory inode at block %"PRIu64"\n", i);
memset(&ds, 0, sizeof(ds));
--- cluster/gfs/gfs_fsck/pass3.c 2005/03/03 19:53:57 1.8
+++ cluster/gfs/gfs_fsck/pass3.c 2006/06/14 19:12:07 1.9
@@ -205,6 +205,7 @@
* until we find one that's been checked already. If we don't
* find a parent, put in lost+found.
*/
+ log_info("Checking directory linkage.\n");
for(i = 0; i < FSCK_HASH_SIZE; i++) {
osi_list_foreach(tmp, &sbp->dir_hash[i]) {
di = osi_list_entry(tmp, struct dir_info, list);
@@ -270,8 +271,8 @@
break;
}
else {
- log_info("Directory at block %"PRIu64
- " connected\n", di->dinode);
+ log_debug("Directory at block %" PRIu64 " connected\n",
+ di->dinode);
}
di = tdi;
}
--- cluster/gfs/gfs_fsck/pass4.c 2005/03/07 22:04:58 1.11
+++ cluster/gfs/gfs_fsck/pass4.c 2006/06/14 19:12:07 1.12
@@ -59,8 +59,8 @@
(ii->inode == sbp->sb.sb_quota_di.no_addr) ||
(ii->inode == sbp->sb.sb_license_di.no_addr))
continue;
- log_info("Checking reference count on inode at block %"PRIu64
- "\n", ii->inode);
+ log_debug("Checking reference count on inode at block %"PRIu64
+ "\n", ii->inode);
if(ii->counted_links == 0) {
log_err("Found unlinked inode at %"PRIu64"\n",
ii->inode);
@@ -174,6 +174,7 @@
if(sbp->lf_dip)
log_debug("At beginning of pass4, l+f entries is %u\n",
sbp->lf_dip->i_di.di_entries);
+ log_info("Checking inode reference counts.\n");
for (i = 0; i < FSCK_HASH_SIZE; i++) {
list = &sbp->inode_hash[i];
if(scan_inode_list(sbp, list)) {
--- cluster/gfs/gfs_fsck/pass5.c 2006/04/20 20:37:50 1.9
+++ cluster/gfs/gfs_fsck/pass5.c 2006/06/14 19:12:07 1.10
@@ -17,6 +17,7 @@
#include "ondisk.h"
#include "fs_bits.h"
#include "bio.h"
+#include "util.h"
#ifdef DEBUG
int rgrp_countbits(unsigned char *buffer, unsigned int buflen,
@@ -191,6 +192,8 @@
while(byte < end) {
rg_status = ((*byte >> bit) & GFS_BIT_MASK);
block = rg_data + *rg_block;
+ log_debug("Checking block %" PRIu64 "\n", block);
+ warm_fuzzy_stuff(block);
block_check(sbp->bl, block, &q);
block_status = convert_mark(q.block_type, count);
--- cluster/gfs/gfs_fsck/super.c 2006/04/20 20:37:50 1.6
+++ cluster/gfs/gfs_fsck/super.c 2006/06/14 19:12:07 1.7
@@ -1081,9 +1081,9 @@
gfs_grow or something. Count the RGs by hand. */
} trust_lvl;
- log_warn("Validating Resource Group index.\n");
+ log_info("Validating Resource Group index.\n");
for (trust_lvl = blind_faith; trust_lvl <= distrust; trust_lvl++) {
- log_warn("Level %d check.\n", trust_lvl + 1);
+ log_info("Level %d check.\n", trust_lvl + 1);
count1 = count2 = 0;
/* ---------------------------------------------------------------- */
/* Step 1 - Calculate or figure out our own RG index */
@@ -1097,7 +1097,7 @@
error = gfs_rgindex_calculate(sdp, &expected_rglist,
&calc_rg_count);
if (error) { /* If calculated RGs don't reasonably match the fs */
- log_warn("(failed--trying again at level 3)\n");
+ log_info("(failed--trying again@level 3)\n");
ri_cleanup(&sdp->rglist);
continue; /* Try again, this time counting them manually */
}
@@ -1106,7 +1106,7 @@
error = gfs_rgindex_rebuild(sdp, &expected_rglist,
&calc_rg_count); /* count the RGs. */
if (error) { /* If calculated RGs don't reasonably match the fs */
- log_err("(failed--giving up)\n");
+ log_info("(failed--giving up)\n");
goto fail; /* try again, this time counting them manually */
}
}
@@ -1177,7 +1177,7 @@
count1++;
} /* for all RGs in the index */
if (!error) {
- log_warn("%u resource groups found.\n", rg);
+ log_info("%u resource groups found.\n", rg);
if (trust_lvl != blind_faith && rg != calc_rg_count)
log_warn("Resource group count discrepancy. Index says %d. " \
"Should be %d.\n", rg, calc_rg_count);
@@ -1205,15 +1205,15 @@
sdp->rgcount = count1;
}
if (!error) { /* if no problems encountered with the rgs */
- log_warn("(passed)\n");
+ log_info("(passed)\n");
break; /* no reason to distrust what we saw. Otherwise, we
reiterate and become a little less trusting. */
}
else {
if (trust_lvl < distrust)
- log_warn("(failed--trying again at level 2)\n");
+ log_info("(failed--trying again@level 2)\n");
else
- log_err("(failed--recovery impossible)\n");
+ log_info("(failed--recovery impossible)\n");
}
ri_cleanup(&sdp->rglist);
} /* for trust_lvl */
--- cluster/gfs/gfs_fsck/util.c 2005/03/03 19:53:57 1.6
+++ cluster/gfs/gfs_fsck/util.c 2006/06/14 19:12:07 1.7
@@ -313,3 +313,24 @@
return NULL;
}
#endif
+
+/* Put out a warm, fuzzy message every second so the user */
+/* doesn't think we hung. (This may take a long time). */
+void warm_fuzzy_stuff(uint64_t block)
+{
+ static struct timeval tv;
+ static uint32_t seconds = 0;
+
+ gettimeofday(&tv, NULL);
+ if (!seconds)
+ seconds = tv.tv_sec;
+ if (tv.tv_sec - seconds) {
+ uint64_t percent;
+
+ seconds = tv.tv_sec;
+ if (last_fs_block) {
+ percent = (block * 100) / last_fs_block;
+ log_notice("\r%" PRIu64 " percent complete.\r", percent);
+ }
+ }
+}
--- cluster/gfs/gfs_fsck/util.h 2005/03/03 19:53:57 1.3
+++ cluster/gfs/gfs_fsck/util.h 2006/06/14 19:12:07 1.4
@@ -35,5 +35,6 @@
int next_rg_meta_free(struct fsck_rgrp *rgd, uint64 *block, int first, int *free);
int next_rg_metatype(struct fsck_rgrp *rgd, uint64 *block, uint32 type, int first);
struct di_info *search_list(osi_list_t *list, uint64 addr);
+void warm_fuzzy_stuff(uint64_t block);
#endif /* __UTIL_H__ */
next reply other threads:[~2006-06-14 19:12 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-06-14 19:12 rpeterso [this message]
-- strict thread matches above, loose matches on Subject: below --
2006-06-20 14:43 [Cluster-devel] cluster/gfs/gfs_fsck fs_recovery.c fsck.h init rpeterso
2006-06-21 14:21 rpeterso
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20060614191209.5876.qmail@sourceware.org \
--to=rpeterso@sourceware.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.