[Stgt-devel] Bug report - create a file on mounted FS (ext3) with dd fails

[Eli Dorfman]

n Mon, Jul 21, 2008 at 4:05 PM, Pete Wyckoff <pw at osc.edu> wrote:
> dorfman.eli at gmail.com wrote on Mon, 21 Jul 2008 10:28 +0300:
>> I have the following test environment:
>> Initiator: RH4.4 with OFED1.3.1
>> Target: RH5.1 with Stgt latest snapshot 20080715
>> I'm using iser transport
>>
>> I created an ext3 FS on the device
>> #mkfs.ext3 /dev/sdc
>>
>> mounted the device
>> # mount /dev/sdc /mnt/sdc
>>
>> Trying to create a file on the mounted FS
>> # dd if=/dev/zero of=/mnt/sdc/tempfile bs=1k count=40M
>>
>> I got the following error on the initiator:
>>
>> dd: writing `/mnt/sdc/tempfile': Read-only file system
>> 1489+0 records in
>> 1488+0 records out
>
> Take a look at the initiator logs to see if there is anything fishy.

Already did. It shows:
Jul 21 11:54:10 seed5 kernel: EXT3-fs error (device sdc):
ext3_new_block: Allocating block in system zone - block = 557056
Jul 21 11:54:10 seed5 kernel: Aborting journal on device sdc.
Jul 21 11:54:10 seed5 kernel: ext3_abort called.

This problem does not happen
with iser transport with ext2 OR
with tcp transport with ext3

It seems to be related to ext3.

>
>> And the following messages on the target:
>> tgtd: iser_rx_progress(1101) entry
>> tgtd: handle_wc(880) outgoing rsp complete
>> tgtd: handle_wc(880) outgoing rsp complete
>> tgtd: iser_rx_progress(1101) entry
>> bs_rdwr_request(121) io done 0x546df8 2a 4096 4096
>> tgtd: bs_thread_ack_fn(82) found 0x546df8
>> tgtd: bs_thread_request_done(122) back to tgtd, 0x546df8
>> tgtd: iscsi_rdma_event_modify(1628) tx ready adding 0x5327e0
>> tgtd: iser_tx_progress(1017) entry
>> tgtd: iser_tx_progress(1028) trying tx
>> tgtd: iscsi_task_tx_start(1745) found a task 7b 4096 0 0
>> tgtd: iscsi_iser_write_begin(1352) new sendl 0x2aaaba9fd090 len 48
>> tgtd: iscsi_iser_write_end(1387) sendl 0x2aaaba9fd090 len 76
>> tgtd: iscsi_iser_write_end(1402) inc progress to finish cmd
>> tgtd: __cmd_done(894) 8 0x2aaaac22c000 (nil) 4096 0 0
>> tgtd: iscsi_rdma_free_data_buf(1678) free 0x2aaaac22c000
>> tgtd: iser_tx_progress(1028) trying tx
>> tgtd: iscsi_task_tx_start(1770) no more data
>> tgtd: iscsi_rdma_event_modify(1633) tx ready removing 0x5327e0
>> tgtd: iser_tx_progress(1017) entry
>> tgtd: iser_rx_progress(1101) entry
>> tgtd: handle_wc(880) outgoing rsp complete
>
> This is all normal.  A 4 kB write completed with no errors.

That's what I thought.

Thanks,
Eli