Leif Hedstrom created TS-3377:
---------------------------------
Summary: Some disks in storage.config considered bad for no good
reason
Key: TS-3377
URL: https://issues.apache.org/jira/browse/TS-3377
Project: Traffic Server
Issue Type: Bug
Components: Cache
Reporter: Leif Hedstrom
Using latest master, I have a config like
{code}
/dev/xvdb1
/dev/xvdb2
{code}
The files are properly udev'ed, i.e.
{code}
[root@qa1 ats]# ls -lrt /dev/xvdb[12]
brw-rw----. 1 root nobody 202, 18 Feb 7 00:36 /dev/xvdb2
brw-rw----. 1 root nobody 202, 17 Feb 7 00:36 /dev/xvdb1
{code}
If I initialize the cache, everything looks fine:
{code}
[root@qa1 ats]# ./bin/traffic_server -Cclear -T cache
traffic_server: using root directory '/opt/ats'
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.ram_cache.size = 1073741824 = 1024Mb
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.limits.http.max_alts = 4
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
cache_config_ram_cache_cutoff = 4194304 = 4Mb
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.permit.pinning = 0
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.dir.sync_frequency = 600
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.select_alternate = 1
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.max_doc_size = 0 = 0Mb
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.mutex_retry_delay = 2ms
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.hit_evacuate_percent = 0
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.hit_evacuate_size_limit = 0
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.max_disk_errors = 5
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.agg_write_backlog = 5242880
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.enable_checksum = 0
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.alt_rewrite_max_size = 4096
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
proxy.config.cache.enable_read_while_writer = 1
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
Store::read_config, fd = -1, "/opt/ats/etc/trafficserver/storage.config"
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
Store::read_config: "/dev/xvdb1"
[Feb 7 00:36:33.588] Server {0x7fc313646880} DEBUG: (cache_init)
Store::read_config - ns = new Span; ns->init("/dev/xvdb1",-1), forced volume=-1
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init) initialized
span '/dev/xvdb1'
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init)
hw_sector_size=512, size=23040000000, blocks=2812500, disk_id=0/51729,
file_pathname=1
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init)
Store::read_config: "/dev/xvdb2"
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init)
Store::read_config - ns = new Span; ns->init("/dev/xvdb2",-1), forced volume=-1
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init) initialized
span '/dev/xvdb2'
[Feb 7 00:36:33.589] Server {0x7fc313646880} DEBUG: (cache_init)
hw_sector_size=512, size=23040000000, blocks=2812500, disk_id=0/51730,
file_pathname=1
[Feb 7 00:36:33.601] Server {0x7fc313646880} DEBUG: (cache_init) initialized
span '/opt/ats/var/trafficserver'
[Feb 7 00:36:33.601] Server {0x7fc313646880} DEBUG: (cache_init)
hw_sector_size=4096, size=1048576, blocks=128, disk_id=51713/1710129,
file_pathname=0
[Feb 7 00:36:33.603] Server {0x7fc313646880} DEBUG: (cache_hosting) Disk: 0,
blocks: 2812500
[Feb 7 00:36:33.609] Server {0x7fc313646880} DEBUG: (cache_hosting) Disk: 1,
blocks: 2812500
[Feb 7 00:36:33.617] Server {0x7fc309824700} DEBUG: (cache_hosting) blocks =
2812498
[Feb 7 00:36:33.617] Server {0x7fc309824700} DEBUG: (cache_hosting) Disk: 0:
Vol Blocks: 1: Free space: 0
[Feb 7 00:36:33.617] Server {0x7fc309824700} DEBUG: (cache_hosting) Vol: 0
Size: 2812498
[Feb 7 00:36:33.617] Server {0x7fc309824700} DEBUG: (cache_hosting) Block
No: 0 Size: 2812498 Free: 0
[Feb 7 00:36:33.617] Server {0x7fc309824700} DEBUG: (cache_init) Cache::open -
proxy.config.cache.min_average_object_size = 8000
[Feb 7 00:36:33.619] Server {0x7fc309824700} DEBUG: (cache_init) allocating
28745728 directory bytes for a 23039983616 byte volume (0.124765%)
[Feb 7 00:36:33.781] Server {0x7fc308c0f700} DEBUG: (cache_hosting) Host
Record: 0x60b000015fa8, Volume: 0, size: 2812498
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
build_vol_hash_table index 0 mapped to 0 requested 32707 got 32707
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - theCache, total_size = 2812498 = 21972 MB
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - caches_ready=0x9, gnvol=1
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - 1073741824 != AUTO_SIZE_RAM_CACHE
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - http_ram_cache_size = 1073741824 = 1024Mb
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - stream_ram_cache_size = 0 = 0Mb
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - factor = 1.000000
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized[0] - ram_cache_bytes = 1073741824 = 1024Mb
[Feb 7 00:36:33.782] Server {0x7fc308c0f700} DEBUG: (cache_init)
CacheProcessor::cacheInitialized - total_cache_bytes = 23011237888 = 21945Mb
{code}
However, when I start up ATS properly, I get the following errors in digs.log:
{code}
[Feb 7 00:39:01.825] {0x2b44b022d1c0} STATUS: opened
/opt/ats/var/log/trafficserver/diags.log
[Feb 7 00:39:01.825] {0x2b44b022d1c0} NOTE: updated diags config
[Feb 7 00:39:01.835] Server {0x2b44b022d1c0} NOTE: cache clustering disabled
[Feb 7 00:39:01.846] Server {0x2b44b022d1c0} NOTE: ip_allow.config updated,
reloading
[Feb 7 00:39:01.854] Server {0x2b44b022d1c0} NOTE: cache clustering disabled
[Feb 7 00:39:01.856] Server {0x2b44b022d1c0} NOTE: logging initialized[3],
logging_mode = 3
[Feb 7 00:39:01.857] Server {0x2b44b022d1c0} NOTE: loading plugin
'/opt/ats/libexec/trafficserver/stats_over_http.so'
[Feb 7 00:39:01.860] Server {0x2b44b022d1c0} NOTE: loading SSL certificate
configuration from /opt/ats/etc/trafficserver/ssl_multicert.config
[Feb 7 00:39:01.877] Server {0x2b44beabc700} WARNING: cache disk operation
failed READ -1 21
[Feb 7 00:39:01.884] Server {0x2b44b022d1c0} NOTE: traffic server running
[Feb 7 00:39:01.884] Server {0x2b44ba422700} WARNING: could not read disk
header for disk /dev/xvdb2: declaring disk bad
[Feb 7 00:39:01.977] Server {0x2b44b9c14700} NOTE: cache enabled
{code}
I've tried this with 3 and 4 disks as well, and it always fails to setup
anything but the first disk at normal startup.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
